CN114730521A - Collision Detection Using Statistical Models - Google Patents

Abstract

Techniques and methods for performing collision monitoring using an error model. For example, a vehicle may generate sensor data using one or more sensors. The vehicle may then use the system to analyze the sensor data to determine a parameter associated with the vehicle and a parameter associated with another object. Further, the vehicle may process parameters associated with the vehicle using an error model associated with the system to determine a distribution of estimated locations associated with the vehicle. The vehicle may also process parameters associated with the object using an error model to determine a distribution of estimated locations associated with the object. Using the distribution of estimated locations, the vehicle can determine a probability of collision between the vehicle and the object.

Description

Translated fromChinese

使用统计模型的碰撞监测Collision Detection Using Statistical Models

相关申请的交叉引用CROSS-REFERENCE TO RELATED APPLICATIONS

本申请请求以于2019年11月13日提交的题为“使用统计模型的碰撞监测(COLLISION MONITORING USING STATISTIC MODELS)”的美国专利申请第16/682,971号、以及于2019年11月13日提交的题为“使用系统数据的碰撞监测(COLLISION MONITORINGUSING SYSTEM DATA)”的美国专利申请第16/683,005号为优先权，其全部内容通过引用并入本文。This application requests US Patent Application Serial No. 16/682,971, entitled "COLLISION MONITORING USING STATISTIC MODELS", filed on November 13, 2019, and filed on November 13, 2019 US Patent Application No. 16/683,005 entitled "COLLISION MONITORINGUSING SYSTEM DATA" is of priority and is incorporated herein by reference in its entirety.

背景技术Background technique

自动驾驶车辆可以使用自动驾驶车辆控制器来引导自动驾驶车辆通过环境。例如，自动驾驶车辆控制器可以使用规划方法、装置和系统来确定行驶路径并引导自动驾驶车辆通过环境，该环境包含动态物体(例如，车辆、行人、动物等)和静态物体(例如，建筑物、招牌、停止的车辆等)。当车辆导航通过环境时，自动驾驶车辆控制器可以将动态物体的预测行为纳入考虑。An autonomous vehicle may use an autonomous vehicle controller to guide the autonomous vehicle through the environment. For example, an autonomous vehicle controller may use planning methods, devices, and systems to determine a driving path and guide the autonomous vehicle through an environment that includes both dynamic objects (eg, vehicles, pedestrians, animals, etc.) and static objects (eg, buildings) , signs, stopped vehicles, etc.). Autonomous vehicle controllers can take the predicted behavior of dynamic objects into account as the vehicle navigates through the environment.

附图说明Description of drawings

以下详细描述是参照附图来加以叙述的。在附图中，附图标记的最左侧数字标识第一次出现该附图标记的附图。在不同附图中使用相同的附图标记表示相似或相同的组件或特征。The following detailed description is set forth with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The use of the same reference numbers in different drawings indicates similar or identical components or features.

图1是根据本公开的实施例的包括使用误差模型和/或系统数据执行碰撞监测的车辆的环境的示图。1 is an illustration of an environment including a vehicle performing collision monitoring using error models and/or system data, in accordance with an embodiment of the present disclosure.

图2是根据本公开的实施例的车辆使用误差模型分析传感器数据以便确定与物体相关联的估计位置的示例的示图。2 is a diagram of an example of a vehicle analyzing sensor data using an error model to determine an estimated location associated with an object, in accordance with an embodiment of the present disclosure.

图3是根据本公开的实施例的车辆使用误差模型分析传感器数据以便确定与物体相关联的估计位置的另一示例的示图。3 is a diagram of another example of a vehicle analyzing sensor data using an error model to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure.

图4是根据本公开的实施例的车辆分析传感器数据和系统数据以便确定与物体相关联的估计位置的示例的示图。4 is a diagram of an example of a vehicle analyzing sensor data and system data in order to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure.

图5是根据本公开的实施例的车辆分析传感器数据和系统数据以便确定与物体相关联的估计位置的另一示例的示图。5 is a diagram of another example of a vehicle analyzing sensor data and system data in order to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure.

图6示出了根据本公开的实施例的说明车辆确定碰撞概率的示例图表。FIG. 6 shows an example graph illustrating vehicle determination of collision probability in accordance with an embodiment of the present disclosure.

图7示出了根据本公开的实施例的至少部分地基于车辆数据和地面真值数据生成误差模型数据。7 illustrates generating error model data based at least in part on vehicle data and ground truth data in accordance with an embodiment of the present disclosure.

图8示出了根据本公开的实施例的至少部分地基于由一个或多个车辆生成的记录数据和地面真值数据生成感知误差模型数据的一个或多个计算设备。8 illustrates one or more computing devices generating perceptual error model data based at least in part on recorded data and ground truth data generated by one or more vehicles, in accordance with embodiments of the present disclosure.

图9示出了根据本公开的实施例的至少部分地基于车辆数据和地面真值数据生成不确定性数据。9 illustrates generating uncertainty data based at least in part on vehicle data and ground truth data in accordance with an embodiment of the present disclosure.

图10描绘了根据本公开的实施例的用于实现本文描述的技术的示例系统的方框图。10 depicts a block diagram of an example system for implementing the techniques described herein in accordance with embodiments of the present disclosure.

图11描绘了根据本公开的实施例的用于使用误差模型执行碰撞监测的示例过程。11 depicts an example process for performing collision monitoring using an error model, according to an embodiment of the present disclosure.

图12描绘了根据本公开的实施例的用于使用误差模型来确定与物体相关联的估计位置的示例过程。12 depicts an example process for determining an estimated location associated with an object using an error model, according to an embodiment of the present disclosure.

图13A至图13B描绘了根据本公开的实施例的用于使用不确定性执行碰撞监测的示例过程。13A-13B depict an example process for performing collision monitoring using uncertainty, according to embodiments of the present disclosure.

图14描绘了根据本公开的实施例的用于使用不确定性来确定与物体相关联的估计位置的示例过程。14 depicts an example process for using uncertainty to determine an estimated location associated with an object, according to an embodiment of the present disclosure.

具体实施方式Detailed ways

如上所述，自动驾驶车辆可以使用控制器来引导自动驾驶车辆通过环境。例如，控制器可以使用规划方法、设备和系统来确定行驶路径并引导自动驾驶车辆通过环境，该环境包括动态物体(例如，车辆、行人、动物等)和/或静态物体(例如，建筑物、招牌、停止的车辆等)。为了确保乘员和物体的安全，自动驾驶车辆控制器在环境中运行时可以采用安全因子。然而，在至少一些示例中，这样的系统和控制器可以包括无法被检查的复杂系统。尽管存在这样的事实：可能没有用于确定与此类系统和系统相关联的误差或不确定性的方法，此类误差和不确定性对于告知此类车辆在环境中的安全操作可能是必要的。As mentioned above, an autonomous vehicle can use a controller to guide the autonomous vehicle through an environment. For example, a controller may use planning methods, devices, and systems to determine a driving path and guide an autonomous vehicle through an environment that includes dynamic objects (eg, vehicles, pedestrians, animals, etc.) and/or static objects (eg, buildings, signs, stopped vehicles, etc.). To ensure the safety of occupants and objects, autonomous vehicle controllers can employ safety factors when operating in the environment. However, in at least some examples, such systems and controllers may include complex systems that cannot be inspected. Notwithstanding the fact that there may be no method for determining the errors or uncertainties associated with such systems and systems, such errors and uncertainties may be necessary to inform the safe operation of such vehicles in the environment .

因此，本公开针对这样的技术：用于通过确定复杂系统和系统的此类误差和/或不确定性模型来使用误差模型和/或系统数据执行碰撞监测。例如，自动驾驶车辆可以使用误差模型和/或系统不确定性来在以后的时间确定自动驾驶车辆和一个或多个物体两者的估计位置。在一些情况下，估计位置可以包括与自动驾驶车辆和一个或多个物体相关联的概率位置的分布。然后，自动驾驶车辆可以使用估计位置确定自动驾驶车辆与一个或多个物体之间的碰撞概率。至少部分地基于碰撞概率，自动驾驶车辆可以执行一个或多个动作。在至少一些示例中，可以基于根据本文所详细描述的任何技术做出的确定来确定这样的概率。Accordingly, the present disclosure is directed to techniques for performing collision monitoring using error models and/or system data by determining such error and/or uncertainty models for complex systems and systems. For example, the self-driving vehicle may use the error model and/or system uncertainty to determine the estimated location of both the self-driving vehicle and one or more objects at a later time. In some cases, the estimated location may include a distribution of probabilistic locations associated with the autonomous vehicle and one or more objects. The autonomous vehicle can then use the estimated location to determine the probability of a collision between the autonomous vehicle and one or more objects. Based at least in part on the probability of collision, the autonomous vehicle may perform one or more actions. In at least some examples, such probabilities may be determined based on determinations made in accordance with any of the techniques described in detail herein.

有关更多详细信息，自动驾驶车辆可以穿越环境并使用一个或多个传感器生成传感器数据。在一些情况下，传感器数据可以包括由传感器捕获的数据，该传感器例如为飞行时间传感器、位置传感器(例如，GPS、指南针等)、惯性传感器(例如，惯性测量单元(IMU)、加速度计、磁力计、陀螺仪等)、激光雷达传感器、雷达传感器、声纳传感器、红外传感器、摄像机(例如，RGB、IR、强度、深度等)、麦克风传感器、环境传感器(例如，温度传感器、湿度传感器、光传感器、压力传感器等)、超声波转换器、车轮编码器等。然后，自动驾驶车辆可以在导航通过环境时使用一个或多个组件(例如，一个或多个系统)来分析传感器数据。For more details, an autonomous vehicle can traverse the environment and generate sensor data using one or more sensors. In some cases, sensor data may include data captured by sensors such as time-of-flight sensors, location sensors (eg, GPS, compass, etc.), inertial sensors (eg, inertial measurement units (IMUs), accelerometers, magnetic meters, gyroscopes, etc.), lidar sensors, radar sensors, sonar sensors, infrared sensors, cameras (eg, RGB, IR, intensity, depth, etc.), microphone sensors, environmental sensors (eg, temperature sensors, humidity sensors, light sensors, pressure sensors, etc.), ultrasonic transducers, wheel encoders, etc. The autonomous vehicle can then use one or more components (eg, one or more systems) to analyze the sensor data as it navigates through the environment.

例如，自动驾驶车辆的组件中的一个或多个组件可以使用传感器数据来生成自动驾驶车辆的轨迹。在一些情况下，一个或多个组件还可以使用传感器数据来确定与自动驾驶车辆的位置相关联的姿势数据。例如，一个或多个组件可以使用传感器数据来确定车辆在环境中的位置数据、坐标数据和/或方位数据。在一些情况下，姿势数据可以包括x-y-z坐标和/或可以包括与车辆相关联的俯仰(pitch)、滚动(roll)和偏航(yaw)数据。For example, one or more of the components of an autonomous vehicle may use sensor data to generate a trajectory of the autonomous vehicle. In some cases, one or more components may also use sensor data to determine gesture data associated with the location of the autonomous vehicle. For example, one or more components may use sensor data to determine position data, coordinate data, and/or orientation data of the vehicle in the environment. In some cases, the gesture data may include x-y-z coordinates and/or may include pitch, roll, and yaw data associated with the vehicle.

此外，自动驾驶车辆的一个或多个组件可以使用传感器数据来执行操作，例如，检测、识别、分割、分类和/或追踪环境内的物体。例如，诸如行人、自行车/自行车骑士、摩托车/摩托车骑士、公共汽车、有轨电车、卡车、动物和/或其类似物的物体可以存在于环境中。一个或多个组件可以使用传感器数据来确定物体的当前位置以及物体在未来时间(例如，未来一秒、未来五秒等)的估计位置。Additionally, one or more components of the autonomous vehicle may use the sensor data to perform operations, such as detecting, identifying, segmenting, classifying, and/or tracking objects within the environment. For example, objects such as pedestrians, bicycles/cyclists, motorcycles/motorcyclists, buses, streetcars, trucks, animals and/or the like may be present in the environment. One or more components may use the sensor data to determine the current location of the object and the estimated location of the object at a future time (eg, one second in the future, five seconds in the future, etc.).

然后，自动驾驶车辆可以使用自动驾驶车辆的轨迹以及物体的估计位置来确定自动驾驶车辆和物体之间的碰撞概率。例如，自动驾驶车辆可以确定物体在未来时间的估计位置是否与自动驾驶车辆在未来时间沿着轨迹的位置相交。为了提高安全性，自动驾驶车辆可以在做出决定时使用距离和/或时间缓冲。例如，当物体在未来时间的位置在与自动驾驶车辆的位置的阈值距离(例如，距离缓冲)内时，自动驾驶车辆可以确定存在着高碰撞概率。The self-driving vehicle can then use the self-driving vehicle's trajectory and the object's estimated location to determine the probability of a collision between the self-driving vehicle and the object. For example, the self-driving vehicle can determine whether the estimated location of an object at a future time intersects the location of the self-driving vehicle along the trajectory at a future time. To improve safety, autonomous vehicles can use distance and/or time buffers when making decisions. For example, the autonomous vehicle may determine that there is a high probability of collision when the location of the object at a future time is within a threshold distance (eg, a distance buffer) from the location of the autonomous vehicle.

此外，自动驾驶车辆可以使用与组件相关联的误差模型和/或与组件的输出相关联的不确定性来确定碰撞概率。与组件相关联的误差模型可以表示与组件的输出相关联的一个或多个误差和/或误差百分比。例如，感知误差模型可以产生与感知组件的感知参数(例如，输出)相关联的感知误差，预测误差模型可以产生与来自预测组件的预测参数(例如，输出)相关联的预测误差，等等。在一些情况下，误差可以通过但不限于查找表来表示，该查找表是使用地面真值数据、函数(例如，基于输入参数的误差)或将输出映射到特定误差的任何其他模型或数据结构至少部分地基于统计聚合来确定的。在至少一些示例中，这样的误差模型可以将特定误差与发生的概率/频率进行映射。如将描述的，在一些示例中，可以为某些类别的数据确定这样的误差模型(例如，感知系统的不同误差模型，其用于基于车辆、物体等的速率在第一范围内的检测和在第二距离范围内的检测)。Additionally, the autonomous vehicle may use error models associated with the components and/or uncertainties associated with the outputs of the components to determine the probability of collision. The error model associated with the component may represent one or more errors and/or error percentages associated with the output of the component. For example, a perceptual error model can generate perceptual errors associated with perceptual parameters (eg, outputs) of a perceptual component, a predictive error model can generate predictive errors associated with predictive parameters (eg, outputs) from a predictive component, and so on. In some cases, the error can be represented by, but is not limited to, a lookup table using ground truth data, a function (eg, error based on input parameters), or any other model or data structure that maps the output to a specific error Determined based at least in part on statistical aggregation. In at least some examples, such an error model may map a particular error to the probability/frequency of occurrence. As will be described, in some examples, such error models may be determined for certain classes of data (eg, different error models of the perception system for detection and detection in the first range based on the velocity of vehicles, objects, etc. detection in the second distance range).

在一些情况下，误差模型可能包括静态误差模型。在其他情况下，误差模型可以包括由自动驾驶车辆和/或一个或多个计算设备更新的动态误差模型。例如，一个或多个计算设备可以持续从自动驾驶车辆接收车辆数据。一个或计算设备然后可以使用车辆数据以及地面真值数据来更新误差模型，这将在下面更详细地描述。在更新误差模型之后，一个或多个计算设备可以将更新的误差模型发送到自动驾驶车辆。In some cases, the error model may include a static error model. In other cases, the error model may include a dynamic error model updated by the autonomous vehicle and/or one or more computing devices. For example, one or more computing devices may continuously receive vehicle data from an autonomous vehicle. An or computing device may then use the vehicle data as well as the ground truth data to update the error model, which will be described in more detail below. After updating the error model, the one or more computing devices may send the updated error model to the autonomous vehicle.

组件可以对传感器数据进行分析，并且至少部分地基于该分析，产生可以表示一个或多个参数的输出。然后，误差模型可以指示车辆的组件的输出，例如，与物体相关联的速度，其与误差百分比相关联。例如，组件可以确定物体在环境中的速度是每秒10米。使用误差模型，自动驾驶车辆可以确定误差百分比为X％(例如，20％)，从而导致速度范围为+/-X％(例如，在20％误差百分比的情况下，范围在每秒8米和每秒12米之间)。在一些情况下，速度范围可以与概率分布相关联，例如，高斯分布，其表示范围的一些部分比范围的其他部分具有更高的发生概率。在一些示例中，可以将概率分布分组成多个离散概率。例如，每秒8米和每秒12米可能与5％的概率相关联，每秒9米和每秒11米可能与20％的概率相关联，而每秒10米可能与45％的概率相关联。The component can analyze the sensor data and, based at least in part on the analysis, generate an output that can represent one or more parameters. The error model can then indicate the output of components of the vehicle, eg, the velocity associated with the object, which is associated with the error percentage. For example, a component can determine that the speed of an object in the environment is 10 meters per second. Using the error model, the autonomous vehicle can determine that the error percentage is X% (eg, 20%), resulting in a speed range of +/-X% (eg, with a 20% error percentage, the range is 8 meters per second and between 12 meters per second). In some cases, the velocity range may be associated with a probability distribution, eg, a Gaussian distribution, which indicates that some parts of the range have a higher probability of occurrence than other parts of the range. In some examples, probability distributions can be grouped into discrete probabilities. For example, 8 meters per second and 12 meters per second may be associated with a 5% probability, 9 meters per second and 11 meters per second may be associated with a 20% probability, and 10 meters per second may be associated with a 45% probability link.

为了使用误差模型，自动驾驶车辆可以至少部分地基于来自组件的输出和与组件相关联的误差模型来确定物体在未来时间的估计位置。估计位置可以对应于位置的概率分布，例如，高斯分布。在一些情况下，自动驾驶车辆通过最初确定与组件和/或参数中的每一个相关联的相应概率分布来确定物体的估计位置。然后，自动驾驶车辆可以使用所有组件和/或参数的概率分布来确定估计位置。例如，自动驾驶车辆可以聚合或组合所有组件和/或参数的概率分布以确定估计位置。聚合和/或组合概率分布可以包括将概率分布相乘，对概率分布求和，和/或将一个或多个其他公式应用于概率分布。To use the error model, the autonomous vehicle may determine an estimated location of the object at a future time based at least in part on the output from the component and the error model associated with the component. The estimated location may correspond to a probability distribution of locations, eg, a Gaussian distribution. In some cases, the autonomous vehicle determines the estimated location of the object by initially determining a respective probability distribution associated with each of the components and/or parameters. The autonomous vehicle can then use the probability distribution of all components and/or parameters to determine the estimated location. For example, an autonomous vehicle may aggregate or combine probability distributions of all components and/or parameters to determine an estimated location. Aggregating and/or combining probability distributions may include multiplying probability distributions, summing probability distributions, and/or applying one or more other formulas to probability distributions.

另外或可替代地，在一些情况下，自动驾驶车辆可以首先使用来自组件的输出来确定与物体相关联的初始估计位置。然后，自动驾驶车辆使用误差模型来确定输出中的每个输出的总误差。自动驾驶车辆可以通过聚合和/或组合来自组件的误差模型中的每个误差模型的误差来确定总误差。接下来，自动驾驶车辆可以使用总误差和初始估计位置来确定估计位置。在这种情况下，估计位置可以包括围绕初始估计位置的可能位置的分布。Additionally or alternatively, in some cases, the autonomous vehicle may first use the output from the component to determine an initial estimated position associated with the object. The self-driving vehicle then uses the error model to determine the total error for each of the outputs. The autonomous vehicle may determine the total error by aggregating and/or combining the errors from each of the component's error models. Next, the autonomous vehicle can use the total error and the initial estimated position to determine the estimated position. In this case, the estimated location may include a distribution of possible locations around the initial estimated location.

对于第一示例，自动驾驶车辆可以使用一个或多个组件来分析传感器数据以便确定与物体相关联的参数。参数可以包括但不限于物体的类型、物体的当前位置、物体的速度、物体的行进方向等。使用误差模型，自动驾驶车辆然后可以确定与物体的类型相关联的概率分布、与物体的当前位置相关联的概率分布、与物体的速度相关联的概率分布、与物体的行进方向相关联的概率分布等。然后，自动驾驶车辆可以使用参数的概率分布来确定物体在未来时间的估计位置。在该第一示例中，估计位置中的每一个(或任何一个或多个)可以表示为位置的概率分布。For a first example, an autonomous vehicle may use one or more components to analyze sensor data in order to determine parameters associated with objects. The parameters may include, but are not limited to, the type of the object, the current position of the object, the speed of the object, the direction of travel of the object, and the like. Using the error model, the autonomous vehicle can then determine a probability distribution associated with the type of object, a probability distribution associated with the object's current location, a probability distribution associated with the object's speed, a probability distribution associated with the object's direction of travel distribution, etc. The autonomous vehicle can then use the probability distribution of the parameters to determine the estimated location of the object at future times. In this first example, each (or any one or more) of the estimated locations may be represented as a probability distribution of locations.

对于第二示例，自动驾驶车辆可以使用一个或多个组件来分析传感器数据，以便再次确定与物体相关联的参数。然后，自动驾驶车辆可以使用这些参数确定物体在未来时间的初始估计位置。此外，在一些示例中，自动驾驶车辆可以使用与参数相关联的误差模型来确定与确定物体的初始估计位置相关联的总误差和/或误差百分比。然后，自动驾驶车辆可以使用初始估计位置和总误差和/或误差百分比来确定物体的估计位置。再次地，在该第二示例中，估计位置中的每一个(或任何一个或多个)可以表示为位置的概率分布。For a second example, the autonomous vehicle may use one or more components to analyze sensor data in order to again determine parameters associated with the object. The self-driving vehicle can then use these parameters to determine the initial estimated location of the object at a future time. Additionally, in some examples, the autonomous vehicle may use the error model associated with the parameters to determine the total error and/or the error percentage associated with determining the initial estimated position of the object. The autonomous vehicle can then use the initial estimated position and the total error and/or error percentage to determine the estimated position of the object. Again, in this second example, each (or any one or more) of the estimated locations may be represented as a probability distribution of locations.

在上述任一示例中，自动驾驶车辆可以使用类似的过程来确定位于环境内的一个或多个其他物体的估计位置。此外，自动驾驶车辆可以使用类似的过程来确定自动驾驶车辆在未来时间的估计位置。例如，自动驾驶车辆可以使用一个或多个组件来分析传感器数据以便确定与自动驾驶车辆相关联的参数。参数可以包括但不限于自动驾驶车辆的位置、自动驾驶车辆的速度、自动驾驶车辆的行进方向等。然后，自动驾驶车辆可以采用与参数相关联的误差模型来确定自动驾驶车辆在未来时间的估计位置。在一些示例中，估计位置中的每一个(或任何一个或多个)可以对应于自动驾驶车辆在未来时间的位置的概率分布。In any of the above examples, an autonomous vehicle may use a similar process to determine the estimated location of one or more other objects located within the environment. Additionally, a similar process can be used by self-driving vehicles to determine the estimated location of the self-driving vehicle at a future time. For example, an autonomous vehicle may use one or more components to analyze sensor data in order to determine parameters associated with the autonomous vehicle. Parameters may include, but are not limited to, the location of the autonomous vehicle, the speed of the autonomous vehicle, the direction of travel of the autonomous vehicle, and the like. The self-driving vehicle can then employ the error model associated with the parameters to determine the estimated location of the self-driving vehicle at future times. In some examples, each (or any one or more) of the estimated locations may correspond to a probability distribution of locations of the autonomous vehicle at future times.

作为使用误差模型来确定自动驾驶车辆和/或物体的估计位置的补充或替代，自动驾驶车辆可以使用与组件和/或输出相关联的系统数据(例如，不确定性模型)来确定估计位置。与参数相关联的不确定性模型可以对应于输出有多少应当被信任的分布和/或系统认为输出的正确程度的度量。例如，如果组件多次分析传感器数据以便确定物体的位置，如果输出包括由地面真值数据指示的位置周围的值的较小分布(例如，在第一范围内)，则该组件将输出低不确定性。此外，如果输出包括由地面真值数据指示的位置周围的值的较大分布(例如，在大于第一范围的第二范围内)，则该组件将输出大的不确定性。自动驾驶车辆可以使用不确定性模型来确定物体在未来时间的估计位置。In addition to or instead of using the error model to determine the estimated location of the autonomous vehicle and/or object, the autonomous vehicle may use system data (eg, uncertainty models) associated with the components and/or outputs to determine the estimated location. The uncertainty model associated with the parameter may correspond to a distribution of how much of the output should be trusted and/or a measure of how correct the system considers the output. For example, if a component analyzes sensor data multiple times to determine the location of an object, the component will output a low value if the output includes a small distribution of values around the location indicated by the ground truth data (eg, within a first range) certainty. Furthermore, this component will output a large uncertainty if the output includes a large distribution of values around the location indicated by the ground truth data (eg, in a second range greater than the first range). Self-driving vehicles can use uncertainty models to determine the estimated locations of objects at future times.

对于第一示例，自动驾驶车辆可以使用一个或多个组件来分析传感器数据以便再次确定与物体相关联的参数。然后，自动驾驶车辆可以确定与确定物体的类型相关联的不确定性模型、与确定物体的当前位置相关联的不确定性模型、与确定物体的速度相关联的不确定性模型、与确定物体的行进方向相关联的不确定性模型等。然后，自动驾驶车辆可以使用与参数相关联的不确定性模型来确定物体在未来时间的估计位置。在该第一示例中，估计位置可以对应于位置的概率分布。For the first example, the autonomous vehicle may use one or more components to analyze sensor data to again determine parameters associated with the object. The autonomous vehicle may then determine an uncertainty model associated with determining the type of the object, an uncertainty model associated with determining the current position of the object, an uncertainty model associated with determining the speed of the object, and an uncertainty model associated with determining the object's current location. The uncertainty model associated with the direction of travel, etc. The autonomous vehicle can then use the uncertainty model associated with the parameters to determine the estimated location of the object at future times. In this first example, the estimated location may correspond to a probability distribution of locations.

对于第二示例，自动驾驶车辆可以使用一个或多个组件来分析传感器数据，以便再次确定与物体相关联的参数。然后，自动驾驶车辆可以使用这些参数确定物体在未来时间的初始估计位置。此外，自动驾驶车辆可以使用与组件确定参数和估计位置相关联的不确定性模型来确定物体在未来时间的估计位置。再次地，在该第二示例中，估计位置可以对应于位置的概率分布。For a second example, the autonomous vehicle may use one or more components to analyze sensor data in order to again determine parameters associated with the object. The self-driving vehicle can then use these parameters to determine the initial estimated location of the object at a future time. Additionally, the autonomous vehicle can use an uncertainty model associated with the components to determine the parameters and the estimated location to determine the estimated location of the object at a future time. Again, in this second example, the estimated location may correspond to a probability distribution of locations.

在上述任一示例中，自动驾驶车辆可以使用类似的过程来确定位于环境内的一个或多个其他物体的估计位置。此外，自动驾驶车辆可以使用类似的过程来确定自动驾驶车辆在未来时间的估计位置。例如，自动驾驶车辆可以使用一个或多个组件来分析传感器数据以便确定与自动驾驶车辆相关联的参数。参数可以包括但不限于自动驾驶车辆的位置、自动驾驶车辆的速度、自动驾驶车辆的行进方向等(例如，任何和/或所有的这些参数都可以从规划系统的输出轨迹中得出)。自动驾驶车辆然后可以使用与确定参数相关联的不确定性模型来确定自动驾驶车辆在未来时间的估计位置。在一些示例中，估计位置可以对应于自动驾驶车辆在未来时间的位置的概率分布。In any of the above examples, an autonomous vehicle may use a similar process to determine the estimated location of one or more other objects located within the environment. Additionally, a similar process can be used by self-driving vehicles to determine the estimated location of the self-driving vehicle at a future time. For example, an autonomous vehicle may use one or more components to analyze sensor data in order to determine parameters associated with the autonomous vehicle. Parameters may include, but are not limited to, the position of the autonomous vehicle, the speed of the autonomous vehicle, the direction of travel of the autonomous vehicle, etc. (eg, any and/or all of these parameters may be derived from the output trajectory of the planning system). The autonomous vehicle may then use the uncertainty model associated with the determined parameters to determine the estimated location of the autonomous vehicle at a future time. In some examples, the estimated location may correspond to a probability distribution of locations of the autonomous vehicle at future times.

在一些情况下，自动驾驶车辆然后可以使用自动驾驶车辆的估计位置和物体的估计位置来确定碰撞概率。例如，可以使用自动驾驶车辆的估计位置(例如，位置的概率分布)和物体的估计位置(例如，位置的概率分布)之间的几何重叠区域来计算自动驾驶车辆和物体之间的碰撞概率。在一些情况下，如果有多个物体位于环境中，则自动驾驶车辆可以使用针对物体中的每个物体的所确定的碰撞概率来确定与自动驾驶车辆相关联的总碰撞概率。例如，总碰撞概率可以包括针对物体中的每个物体的碰撞概率的总和。In some cases, the autonomous vehicle may then use the estimated location of the autonomous vehicle and the estimated location of the object to determine a probability of collision. For example, the collision probability between the autonomous vehicle and the object may be calculated using the geometric overlap between the estimated position of the autonomous vehicle (eg, a probability distribution of positions) and the estimated position of the object (eg, a probability distribution of positions). In some cases, if multiple objects are located in the environment, the autonomous vehicle may use the determined probability of collision for each of the objects to determine an overall probability of collision associated with the autonomous vehicle. For example, the total collision probability may include the sum of the collision probabilities for each of the objects.

然后，自动驾驶车辆可以确定碰撞概率是否等于或大于阈值(例如，0.5％、1％、5％和/或某个其他阈值百分比)。在一些情况下，如果碰撞概率小于阈值，则自动驾驶车辆可以继续沿着自动驾驶车辆的当前路线导航行进。然而，在一些情况下，如果自动驾驶车辆确定碰撞概率等于或大于阈值，则自动驾驶车辆可以采取一个或多个动作。例如，自动驾驶车辆可以改变自动驾驶车辆的速度(例如，减速)、改变自动驾驶车辆的路线、停放于安全的位置等。The autonomous vehicle may then determine whether the collision probability is equal to or greater than a threshold (eg, 0.5%, 1%, 5%, and/or some other threshold percentage). In some cases, if the probability of collision is less than a threshold, the autonomous vehicle may continue to navigate along the current route of the autonomous vehicle. However, in some cases, the autonomous vehicle may take one or more actions if the autonomous vehicle determines that the probability of collision is equal to or greater than a threshold. For example, the self-driving vehicle may change the speed of the self-driving vehicle (eg, slow down), change the route of the self-driving vehicle, park in a safe location, and the like.

另外或可替代地，在一些情况下，自动驾驶车辆可以至少部分地基于用于确定自动驾驶车辆的估计位置的不确定性模型和用于确定一个或多个物体的估计位置的不确定性模型来确定与对自动驾驶车辆进行导航相关联的总不确定性。然后，自动驾驶车辆可以生成不同的路线并执行类似的过程，以确定与不同路线相关联的总不确定性。此外，自动驾驶车辆可以选择包括最低不确定性的路线。Additionally or alternatively, in some cases, the autonomous vehicle may be based at least in part on an uncertainty model for determining an estimated location of the autonomous vehicle and an uncertainty model for determining an estimated location for one or more objects to determine the total uncertainty associated with navigating an autonomous vehicle. The autonomous vehicle can then generate different routes and perform a similar process to determine the total uncertainty associated with the different routes. Additionally, autonomous vehicles can choose routes that include the lowest uncertainty.

在一些情况下，自动驾驶车辆和/或一个或多个计算设备使用输入数据(例如，记录数据和/或仿真数据)来生成误差模型和/或不确定性模型。例如，自动驾驶车辆和/或一个或多个计算设备可以将输入数据与地面真值数据进行比较。在一些情况下，可以手动标记和/或从其他经过验证的机器学习组件来确定地面真值数据。例如，输入数据可以包括传感器数据和/或由自动驾驶车辆的组件生成的输出数据。自动驾驶车辆和/或一个或多个计算设备可以将输入数据与可以指示环境中的物体的实际参数的地面真值数据进行比较。通过将输入数据与地面真值数据进行比较，自动驾驶车辆和/或一个或多个计算设备可以确定与组件和/或参数相关联的误差和/或不确定性，并且使用该误差生成对应的误差模型和/或使用该不确定性生成对应的不确定性模型。In some cases, the autonomous vehicle and/or one or more computing devices use input data (eg, recorded data and/or simulation data) to generate error models and/or uncertainty models. For example, the autonomous vehicle and/or one or more computing devices may compare the input data to ground truth data. In some cases, ground truth data may be manually labeled and/or determined from other validated machine learning components. For example, input data may include sensor data and/or output data generated by components of the autonomous vehicle. The autonomous vehicle and/or one or more computing devices may compare the input data to ground truth data that may be indicative of actual parameters of objects in the environment. By comparing input data to ground truth data, the autonomous vehicle and/or one or more computing devices can determine errors and/or uncertainties associated with components and/or parameters, and use the errors to generate corresponding error model and/or use the uncertainty to generate a corresponding uncertainty model.

在一些情况下，自动驾驶车辆和/或一个或多个计算设备可以确定与组件相关联的不确定性。例如，自动驾驶车辆和/或一个或多个计算设备可以多次将输入数据输入到组件中，以便从组件接收多个输出(例如，参数)。自动驾驶车辆和/或一个或多个计算设备然后可以分析输出以确定与输出相关联的分布。使用该分布，自动驾驶车辆和/或一个或多个计算设备可以确定不确定性。例如，如果存在大的分布，则自动驾驶车辆和/或一个或多个计算设备可以确定存在大的不确定性。然而，如果存在小的分布，则自动驾驶车辆和/或一个或多个计算设备可以确定存在小的不确定性。In some cases, the autonomous vehicle and/or one or more computing devices may determine uncertainties associated with the components. For example, an autonomous vehicle and/or one or more computing devices may input input data into a component multiple times in order to receive multiple outputs (eg, parameters) from the component. The autonomous vehicle and/or one or more computing devices may then analyze the output to determine a distribution associated with the output. Using this distribution, the autonomous vehicle and/or one or more computing devices can determine uncertainty. For example, if there is a large distribution, the autonomous vehicle and/or one or more computing devices may determine that there is a large uncertainty. However, if there is a small distribution, the autonomous vehicle and/or one or more computing devices may determine that there is a small uncertainty.

本文描述的技术可以以多种方式实现。下文参考以下附图提供示例实施方式。尽管在自动驾驶车辆的上下文中进行了讨论，但本文所描述的方法、装置和系统可以应用于各种系统(例如，传感器系统或机器人平台)，而不限于自动驾驶车辆。在另一示例中，这些技术可以用于航空或航海上下文中，或用在使用机器视觉的任何系统中(例如，在使用图像数据的系统中)。此外，本文描述的技术可以与真实数据(例如，使用一个或多个传感器捕获的)、仿真数据(例如，由仿真器生成的)或两者的任何组合一起使用。The techniques described herein can be implemented in a variety of ways. Example embodiments are provided below with reference to the following figures. Although discussed in the context of autonomous vehicles, the methods, apparatus, and systems described herein can be applied to a variety of systems (eg, sensor systems or robotic platforms) and are not limited to autonomous vehicles. In another example, these techniques may be used in an aviation or nautical context, or in any system that uses machine vision (eg, in a system that uses image data). Furthermore, the techniques described herein can be used with real data (eg, captured using one or more sensors), simulated data (eg, generated by a simulator), or any combination of the two.

图1是根据本公开的实施例的包括使用误差模型和/或系统数据执行碰撞监测的车辆102的环境100的示图。例如，车辆102可以沿着环境100内的轨迹104导航。在导航时，车辆102可以使用车辆102的一个或多个传感器生成传感器数据106并使用车辆102的一个或多个组件108(例如，一个或多个系统)分析传感器数据106。一个或多个组件108可以包括但不限于定位组件、感知组件、预测组件、规划组件等。至少部分地基于该分析，车辆102可以识别位于环境100内的至少第一物体110和第二物体112。FIG. 1 is a diagram of anenvironment 100 including avehicle 102 performing collision monitoring using error models and/or system data, in accordance with an embodiment of the present disclosure. For example,vehicle 102 may navigate alongtrajectory 104 withinenvironment 100 . While navigating, thevehicle 102 may generatesensor data 106 using one or more sensors of thevehicle 102 and analyze thesensor data 106 using one or more components 108 (eg, one or more systems) of thevehicle 102 . One or more components 108 may include, but are not limited to, a positioning component, a perception component, a prediction component, a planning component, and the like. Based at least in part on the analysis, thevehicle 102 may identify at least afirst object 110 and asecond object 112 located within theenvironment 100 .

此外，车辆102可以使用一个或多个组件108分析传感器数据106以便确定在未来时间的与车辆102相关联的估计位置114、与第一物体110相关联的估计位置116、以及与第二物体112相关联的估计位置118。在一些情况下，估计位置114可以包括与车辆102相关联的位置的概率分布，估计位置116可以包括与第一物体110相关联的位置的概率分布，和/或估计位置118可以包括与第二物体112相关联的位置的概率分布。Additionally, thevehicle 102 may analyze thesensor data 106 using one or more components 108 in order to determine an estimatedlocation 114 associated with thevehicle 102 , an estimatedlocation 116 associated with thefirst object 110 , and an estimatedlocation 116 associated with thesecond object 112 at a future time Associated estimatedlocation 118 . In some cases, estimatedlocation 114 may include a probability distribution of locations associated withvehicle 102 , estimatedlocation 116 may include a probability distribution of locations associated withfirst object 110 , and/or estimatedlocation 118 may include a probability distribution of locations associated withsecond object 110 . Probability distribution of locations associated withobjects 112 .

例如，估计位置114可以包括与车辆102相关联的估计位置120(1)、与第一概率相关联的估计位置的第一区域120(2)(例如，第一边界)、与第二概率相关联的估计位置的第二区域120(3)(例如，第二边界)、以及与第三概率相关联的估计位置的第三区域120(4)(例如，第三边界)。在一些情况下，第一概率大于第二概率并且第二概率大于第三概率。例如，车辆102可以确定车辆102将位于估计位置的第一区域120(2)内的概率高于位于可能位置的第二区域120(3)内的概率。另外，车辆102可以确定车辆102将位于估计位置的第二区域120(3)内的概率高于位于估计位置的第三区域120(4)内的概率。For example, estimatedlocation 114 may include estimated location 120(1) associated withvehicle 102, a first region 120(2) of the estimated location associated with a first probability (eg, a first boundary), associated with a second probability A second region 120(3) (eg, a second boundary) of an estimated location associated with a third region 120(4) of an estimated location (eg, a third boundary) associated with a third probability. In some cases, the first probability is greater than the second probability and the second probability is greater than the third probability. For example, thevehicle 102 may determine that the probability that thevehicle 102 will be located within the first area 120(2) of the estimated location is higher than the probability that it will be located within the second area 120(3) of the likely location. Additionally, thevehicle 102 may determine that thevehicle 102 will be located within the second region 120(3) of the estimated location with a higher probability than within the third region 120(4) of the estimated location.

应当注意的是，虽然图1的示例仅示出了估计位置的三个单独的区域，但是在其他示例中，可以存在任意数量的估计位置的区域。此外，更远离估计位置120(1)的区域可以包括比更靠近估计位置120(1)的区域更低的概率。这可以类似地用于物体110和物体112的估计位置中的每一个。It should be noted that while the example of FIG. 1 shows only three separate regions of estimated positions, in other examples there may be any number of regions of estimated positions. Furthermore, regions further away from estimated location 120(1) may include lower probabilities than regions closer to estimated location 120(1). This can be used similarly for each of the estimated positions ofobject 110 andobject 112 .

另外，估计位置116可以包括与第一物体110相关联的估计位置122(1)、与第一概率相关联的估计位置的第一区域122(2)(例如，第一边界)，与第二概率相关联的估计位置的第二区域122(3)(例如，第二边界)、以及与第三概率相关联的估计位置的第三区域122(4)(例如，第三边界)。在一些情况下，第一概率大于第二概率并且第二概率大于第三概率。例如，车辆102可以确定第一物体110将位于估计位置的第一区域122(2)内的概率高于位于可能位置的第二区域122(3)内的概率。此外，车辆102可以确定第一物体110将位于估计位置的第二区域122(3)内的概率高于位于估计位置的第三区域122(4)内的概率。Additionally, the estimatedlocation 116 may include an estimated location 122(1) associated with thefirst object 110, a first region 122(2) (eg, a first boundary) of the estimated location associated with a first probability, and a second A second region 122(3) of estimated locations associated with a probability (eg, a second boundary), and a third region 122(4) of estimated locations associated with a third probability (eg, a third boundary). In some cases, the first probability is greater than the second probability and the second probability is greater than the third probability. For example, thevehicle 102 may determine that thefirst object 110 will be located within the first region 122(2) of the estimated location with a higher probability than within the second region 122(3) of the likely location. Additionally, thevehicle 102 may determine that thefirst object 110 will be located within the second region 122(3) of the estimated location with a higher probability than within the third region 122(4) of the estimated location.

此外，估计位置118可以包括与第二物体112相关联的估计位置124(1)、与第一概率相关联的估计位置的第一区域124(2)(例如，第一边界)、与第二概率相关联的估计位置的第二区域124(3)(例如，第二边界)，以及与第三概率相关联的估计位置的第三区域124(4)(例如，第三边界)。在一些情况下，第一概率大于第二概率并且第二概率大于第三概率。例如，车辆102可以确定第二物体112将位于估计位置的第一区域124(2)内的概率高于位于可能位置的第二区域124(3)内的概率。此外，车辆102可以确定第二物体112将位于估计位置的第二区域124(3)内的概率高于位于估计位置的第三区域124(4)内的概率。Additionally, the estimatedlocation 118 may include an estimated location 124(1) associated with thesecond object 112, a first region 124(2) (eg, a first boundary) of the estimated location associated with a first probability, a A second region 124(3) of estimated locations associated with a probability (eg, a second boundary), and a third region 124(4) of estimated locations associated with a third probability (eg, a third boundary). In some cases, the first probability is greater than the second probability and the second probability is greater than the third probability. For example, thevehicle 102 may determine that thesecond object 112 will be located within the first region 124(2) of the estimated location with a higher probability than within the second region 124(3) of the likely location. Additionally, thevehicle 102 may determine that thesecond object 112 will be located within the second region 124(3) of the estimated location with a higher probability than within the third region 124(4) of the estimated location.

在一些情况下，车辆102可以使用与一个或多个组件108相关联的一个或多个误差模型126来确定估计位置114-118。例如，对于第一物体110，车辆102可以使用一个或多个组件108分析传感器数据106，以便确定与第一物体110相关联的一个或多个参数128。一个或多个参数128可以包括但不限于第一物体110的类型、第一物体110的当前位置(和/或到第一物体110的距离)、第一物体110的速度等。使用一个或多个误差模型126，车辆102然后可以确定第一物体110的估计位置116。In some cases, thevehicle 102 may use one ormore error models 126 associated with one or more components 108 to determine estimated locations 114 - 118 . For example, for thefirst object 110 , thevehicle 102 may analyze thesensor data 106 using one or more components 108 in order to determine one ormore parameters 128 associated with thefirst object 110 . The one ormore parameters 128 may include, but are not limited to, the type of thefirst object 110, the current location of the first object 110 (and/or distance to the first object 110), the speed of thefirst object 110, and the like. Using the one ormore error models 126 , thevehicle 102 may then determine the estimatedlocation 116 of thefirst object 110 .

对于第一示例，车辆102可以使用第一误差模型126来确定与第一物体110的类型相关联的概率分布、使用第二误差模型126来确定与第一物体110的当前位置相关联的概率分布、使用第三误差模型126来确定与第一物体110的速度相关联的概率分布，等等。例如，使用第一物体110的速度，车辆102可以确定第一物体110的速度为每秒1米。车辆102然后可以使用第三误差模型126来确定误差百分比可以为X％(例如，20％)，从而产生速度范围(例如，在20％时，速度在每秒0.8米和每秒1.2米之间)。在一些情况下，误差模型126可以进一步指示范围的一些部分比范围的其他部分具有更高的发生概率。例如，每秒0.8米和每秒1.2米可能与5％的概率相关联，每秒0.9米和每秒1.1米可能与20％的概率相关联，而每秒1米可能与45％的概率相关联。车辆102可以使用类似的过程来确定一个或多个其他参数128的概率分布。For the first example, thevehicle 102 may use thefirst error model 126 to determine the probability distribution associated with the type of thefirst object 110 and thesecond error model 126 to determine the probability distribution associated with the current location of thefirst object 110 , use thethird error model 126 to determine a probability distribution associated with the velocity of thefirst object 110, and so on. For example, using the velocity of thefirst object 110, thevehicle 102 may determine that the velocity of thefirst object 110 is 1 meter per second. Thevehicle 102 may then use thethird error model 126 to determine that the error percentage may be X% (eg, 20%), resulting in a speed range (eg, at 20%, the speed is between 0.8 meters per second and 1.2 meters per second) ). In some cases, theerror model 126 may further indicate that some portions of the range have a higher probability of occurrence than other portions of the range. For example, 0.8 meters per second and 1.2 meters per second may be associated with a 5% probability, 0.9 meters per second and 1.1 meters per second may be associated with a 20% probability, and 1 meter per second may be associated with a 45% probability link. Thevehicle 102 may use a similar process to determine probability distributions for one or moreother parameters 128 .

车辆102然后可以使用参数128的概率分布来确定第一物体110的估计位置116。此外，车辆102可以使用类似的过程来确定车辆102的参数128，确定与车辆102的参数128相关联的概率分布，并且使用概率分布来确定估计位置114。此外，车辆102可以使用类似的过程来确定第二物体112的参数128，确定与第二物体112的参数128相关联的概率分布，并且使用概率分布来确定估计位置116。Thevehicle 102 may then use the probability distribution of theparameters 128 to determine the estimatedlocation 116 of thefirst object 110 . Furthermore, thevehicle 102 may use a similar process to determine theparameters 128 of thevehicle 102 , determine a probability distribution associated with theparameters 128 of thevehicle 102 , and use the probability distribution to determine the estimatedlocation 114 . Additionally, thevehicle 102 may use a similar process to determine theparameters 128 of thesecond object 112 , determine a probability distribution associated with theparameters 128 of thesecond object 112 , and use the probability distribution to determine the estimatedlocation 116 .

对于第二示例，车辆102可以使用第一物体110的参数128以便确定第一物体110的估计位置122(1)。车辆102然后可以使用与用于确定估计位置122(1)的参数128相关联的误差模型126，以便确定参数128的总误差。使用总误差和估计位置122(1)，车辆102可以确定第一物体110的估计位置116。此外，车辆102可以使用类似的过程来确定车辆102的估计位置114和第二物体112的估计位置118。For the second example, thevehicle 102 may use theparameters 128 of thefirst object 110 in order to determine the estimated location 122( 1 ) of thefirst object 110 . Thevehicle 102 may then use theerror model 126 associated with theparameters 128 used to determine the estimated location 122( 1 ) in order to determine the total error for theparameters 128 . Using the total error and estimated position 122( 1 ), thevehicle 102 may determine the estimatedposition 116 of thefirst object 110 . Additionally, thevehicle 102 may use a similar process to determine the estimatedlocation 114 of thevehicle 102 and the estimatedlocation 118 of thesecond object 112 .

作为使用一个或多个误差模型126来确定估计位置114-118的补充或替代，在其他示例中，车辆102可以使用与一个或多个组件108和/或参数128相关联的一个或多个不确定性模型130。例如，来自一个或多个组件108的输出可以包括与确定参数128相关联的一个或多个不确定性模型130。例如，车辆102可以确定与确定第一物体110的类型相关联的第一不确定性模型130、与确定第一物体110的当前位置相关联的第二不确定性模型130、与确定第一物体110的速度相关联的第三不确定性模型130，等等。车辆102然后可以使用参数128和不确定性模型130来确定第一物体110的估计位置116。In addition to or instead of using one ormore error models 126 to determine estimated locations 114 - 118 , in other examples,vehicle 102 may use one or more different parameters associated with one or more components 108 and/orparameters 128 .Deterministic Model 130 . For example, outputs from one or more components 108 may include one ormore uncertainty models 130 associated with determiningparameters 128 . For example, thevehicle 102 may determine afirst uncertainty model 130 associated with determining the type of thefirst object 110 , asecond uncertainty model 130 associated with determining the current location of thefirst object 110 , asecond uncertainty model 130 associated with determining the first object 110 Athird uncertainty model 130 associated with the speed of 110, and so on. Thevehicle 102 may then determine the estimatedlocation 116 of thefirst object 110 using theparameters 128 and theuncertainty model 130 .

对于第一示例，车辆102可以使用第一不确定性模型130来确定与第一物体110的类型相关联的概率分布，使用第二不确定性模型130来确定与第一物体110的当前位置相关联的概率分布，使用第三不确定性模型130来确定与第一物体110的速度相关联的概率分布，等等。例如，使用第一物体110的速度，车辆102可以确定第一物体110的速度是每秒1米。车辆102然后可以确定第一物体的速度的不确定性是20％，因此，确定性是80％。因此，车辆102可以确定速度范围在每秒0.8米和每秒1.2米之间。在一些情况下，车辆102可以进一步确定范围的一些部分比范围的其他部分具有更高的发生概率。例如，每秒0.8米和每秒1.2米可能与5％的概率相关联，每秒0.9米和每秒1.1米可能与20％的概率相关联，而每秒1米可能与45％的概率相关联。车辆102可以使用类似的过程来确定一个或多个其他参数128的概率分布。For the first example, thevehicle 102 may use thefirst uncertainty model 130 to determine a probability distribution associated with the type of thefirst object 110 and use thesecond uncertainty model 130 to determine the probability distribution associated with the current location of thefirst object 110 associated probability distribution, use thethird uncertainty model 130 to determine a probability distribution associated with the velocity of thefirst object 110, and so on. For example, using the velocity of thefirst object 110, thevehicle 102 may determine that the velocity of thefirst object 110 is 1 meter per second. Thevehicle 102 can then determine that the uncertainty of the speed of the first object is 20%, therefore, the certainty is 80%. Thus, thevehicle 102 may determine that the speed range is between 0.8 meters per second and 1.2 meters per second. In some cases, thevehicle 102 may further determine that some portions of the range have a higher probability of occurrence than other portions of the range. For example, 0.8 meters per second and 1.2 meters per second may be associated with a 5% probability, 0.9 meters per second and 1.1 meters per second may be associated with a 20% probability, and 1 meter per second may be associated with a 45% probability link. Thevehicle 102 may use a similar process to determine probability distributions for one or moreother parameters 128 .

然后，车辆102可以使用参数128的概率分布来确定第一物体110的估计位置116。另外，车辆102可以使用类似的过程来确定车辆102的参数128，确定与车辆102的参数128相关联的概率分布，并且使用概率分布来确定估计位置114。此外，车辆102可以使用类似的过程来确定第二物体112的参数128，确定与第二物体112的参数128相关联的概率分布，并且使用概率分布来确定估计位置116。Thevehicle 102 may then determine the estimatedlocation 116 of thefirst object 110 using the probability distribution of theparameters 128 . Additionally, thevehicle 102 may use a similar process to determine theparameters 128 of thevehicle 102 , determine a probability distribution associated with theparameters 128 of thevehicle 102 , and use the probability distribution to determine the estimatedlocation 114 . Additionally, thevehicle 102 may use a similar process to determine theparameters 128 of thesecond object 112 , determine a probability distribution associated with theparameters 128 of thesecond object 112 , and use the probability distribution to determine the estimatedlocation 116 .

对于第二示例，车辆102可以使用第一物体110的参数128以便确定第一物体110的估计位置122(1)。车辆102然后可以使用与参数128相关联的一个或多个不确定性模型130以便确定与估计位置122(1)相关联的总不确定性。使用总不确定性，车辆102可以确定第一物体110的估计位置116。此外，车辆102可以使用类似的过程来确定车辆102的估计位置114和第二物体112的估计位置118。For the second example, thevehicle 102 may use theparameters 128 of thefirst object 110 in order to determine the estimated location 122( 1 ) of thefirst object 110 . Thevehicle 102 may then use the one ormore uncertainty models 130 associated with theparameters 128 in order to determine the overall uncertainty associated with the estimated location 122(1). Using the total uncertainty, thevehicle 102 can determine the estimatedlocation 116 of thefirst object 110 . Additionally, thevehicle 102 may use a similar process to determine the estimatedlocation 114 of thevehicle 102 and the estimatedlocation 118 of thesecond object 112 .

在上述任一示例中，在确定估计位置114-118之后，车辆102可以使用估计位置114-118来确定碰撞概率。例如，车辆102可以确定车辆102和第一物体110之间的碰撞概率。在一些情况下，车辆102可以至少使用车辆102的估计位置114和第一物体110的估计位置116之间的几何重叠区域来确定碰撞概率。In any of the above examples, after determining the estimated locations 114-118, thevehicle 102 may use the estimated locations 114-118 to determine a probability of collision. For example, thevehicle 102 may determine a probability of collision between thevehicle 102 and thefirst object 110 . In some cases, thevehicle 102 may use at least an area of geometric overlap between the estimatedlocation 114 of thevehicle 102 and the estimatedlocation 116 of thefirst object 110 to determine the probability of collision.

更具体地，车辆102的估计位置114可以是具有参数μ_ν，σ_ν的高斯分布(其可以由N(μ_ν,

)表示)。此外，第一物体110的估计位置116可以是具有参数μ₀，σ₀的高斯分布(其可以由N(μ₀,

表示)。估计位置114和估计位置116之间的重叠概率然后可以转化为P[x＝0]，其中x属于

这可以代表与确定重叠概率相关联的一维问题。More specifically, the estimatedlocation 114 of thevehicle 102 may be a Gaussian distribution with parameters_μν ,_σν (which may be defined by N(_μν ,

)express). Furthermore, the estimatedposition 116 of thefirst object 110 may be a Gaussian distribution with parameters μ₀ , σ₀ (which may be represented by N(μ₀ ,

express). The overlap probability between estimatedposition 114 and estimatedposition 116 can then be transformed into P[x=0], where x belongs to

This can represent a one-dimensional problem associated with determining overlap probabilities.

在一些情况下，车辆102可以执行类似的过程以便将一维问题扩展为二维问题。此外，车辆102可以执行类似的过程以便确定车辆102和第二物体112之间的碰撞概率。在一些情况下，车辆102然后可以使用车辆102和第一物体110之间的碰撞概率以及车辆102和第二物体112之间的碰撞概率来确定总碰撞概率。然而，在图1的示例中，由于估计位置114和估计位置118之间没有几何重叠，车辆102和第二物体112之间的碰撞概率可以为零。In some cases, thevehicle 102 may perform a similar process to expand a one-dimensional problem to a two-dimensional problem. Additionally, thevehicle 102 may perform a similar process in order to determine the probability of a collision between thevehicle 102 and thesecond object 112 . In some cases, thevehicle 102 may then use the probability of collision between thevehicle 102 and thefirst object 110 and the probability of collision between thevehicle 102 and thesecond object 112 to determine the overall probability of collision. However, in the example of FIG. 1 , since there is no geometric overlap between estimatedlocation 114 and estimatedlocation 118 , the probability of collision betweenvehicle 102 andsecond object 112 may be zero.

车辆102然后可以确定碰撞概率是否等于或大于阈值。至少部分地基于确定碰撞概率小于阈值，车辆102可以继续沿着轨迹104导航。然而，至少部分地基于确定碰撞概率等于或大于阈值，车辆102可以采取一个或多个动作。一个或多个动作可以包括但不限于沿着新的轨迹导航、改变速度(例如，减速)、停车等。Thevehicle 102 may then determine whether the collision probability is equal to or greater than a threshold. Thevehicle 102 may continue to navigate along thetrajectory 104 based at least in part on determining that the probability of collision is less than the threshold. However, thevehicle 102 may take one or more actions based at least in part on determining that the collision probability is equal to or greater than the threshold. The one or more actions may include, but are not limited to, navigating along a new trajectory, changing speed (eg, slowing down), parking, and the like.

应当注意的是，在一些示例中，车辆102可以执行类似的过程以便确定物体110和物体112之间的碰撞概率。车辆102然后可以至少部分地基于碰撞概率来执行一个或多个动作。例如，如果车辆102确定物体110和物体112之间的碰撞概率等于或大于阈值，则车辆102可以停止。It should be noted that in some examples,vehicle 102 may perform a similar process in order to determine the probability of collision betweenobject 110 andobject 112 . Thevehicle 102 may then perform one or more actions based at least in part on the probability of collision. For example, ifvehicle 102 determines that the probability of collision betweenobject 110 andobject 112 is equal to or greater than a threshold,vehicle 102 may stop.

图2是根据本公开的实施例的车辆102使用一个或多个误差模型126分析传感器数据106以便确定与物体相关联的估计位置的示例的示图。例如，车辆102的一个或多个传感器系统202可以生成传感器数据106。然后可以由车辆102的一个或多个组件108分析传感器数据106。在图2的示例中，一个或多个组件108可以包括定位组件204、感知组件206、规划组件208和预测组件210。然而，在其他示例中，车辆102可以不包括定位组件204、感知组件206、规划组件208或预测组件210中的一个或多个。另外或可替代地，在一些示例中，车辆102可以包括一个或多个附加组件。2 is an illustration of an example of avehicle 102 analyzingsensor data 106 using one ormore error models 126 in order to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure. For example, one ormore sensor systems 202 of thevehicle 102 may generate thesensor data 106 . Thesensor data 106 may then be analyzed by one or more components 108 of thevehicle 102 . In the example of FIG. 2 , the one or more components 108 may include apositioning component 204 , aperception component 206 , aplanning component 208 , and aprediction component 210 . However, in other examples, thevehicle 102 may not include one or more of thepositioning component 204 , theperception component 206 , theplanning component 208 , or theprediction component 210 . Additionally or alternatively, in some examples, thevehicle 102 may include one or more additional components.

然后，组件204-210中的一个或多个可以分析传感器数据106并且至少部分地基于该分析生成输出212-218。在一些情况下，输出212-218可以包括与车辆102和/或物体相关联的参数。对于第一示例，来自定位组件204的输出212可以指示车辆102的位置。对于第二示例，来自感知组件206的输出214可以包括与物体相关联的检测、分割、分类等。对于第三示例，来自规划组件208的输出216可以包括用于车辆102在环境内行进的路径。One or more of the components 204-210 can then analyze thesensor data 106 and generate outputs 212-218 based at least in part on the analysis. In some cases, the outputs 212 - 218 may include parameters associated with thevehicle 102 and/or the object. For the first example, theoutput 212 from thepositioning component 204 may indicate the location of thevehicle 102 . For the second example, theoutput 214 from theperception component 206 may include detection, segmentation, classification, etc. associated with the object. For the third example, theoutput 216 from theplanning component 208 may include a path for thevehicle 102 to travel within the environment.

应当注意的是，虽然在图2的示例中未示出，但是组件204-210中的一个或多个可以使用来自其他组件204-210中的一个或多个的输出212-218以便生成输出212-218。例如，规划组件208可以使用来自定位组件204的输出212以便生成输出216。对于另一示例，规划组件208可以使用来自感知组件206的输出214以便生成输出216。作为使用来自一个或多个其他组件204-210的输出212-218的补充或替代，组件204-210可以使用下文所描述的概率分布220-226。It should be noted that, although not shown in the example of FIG. 2 , one or more of the components 204 - 210 may use the outputs 212 - 218 from one or more of the other components 204 - 210 in order to generate the output 212 -218. For example,planning component 208 can useoutput 212 frompositioning component 204 in order to generateoutput 216 . For another example,planning component 208 can useoutput 214 fromperception component 206 in order to generateoutput 216 . In addition to or in lieu of using the outputs 212-218 from one or more other components 204-210, the components 204-210 may use the probability distributions 220-226 described below.

一个或多个误差组件228可以被配置为使用一个或多个误差模型126来处理输出212-218，以便生成与输出212-218相关联的概率分布220-226。在一些情况下，一个或多个误差组件228可以包含于组件204-210内。例如，定位组件204可以分析传感器数据106，并且至少部分地基于该分析，生成输出212和与输出212相关联的概率分布220两者。对于另一示例，感知组件206可以分析传感器数据106，并且至少部分地基于该分析，生成输出214和与输出214相关联的概率分布222两者。The one ormore error components 228 may be configured to process the outputs 212-218 using the one ormore error models 126 to generate probability distributions 220-226 associated with the outputs 212-218. In some cases, one ormore error components 228 may be included within components 204-210. For example,positioning component 204 can analyzesensor data 106 and, based at least in part on the analysis, generate bothoutput 212 and aprobability distribution 220 associated withoutput 212 . For another example,perception component 206 can analyzesensor data 106 and, based at least in part on the analysis, generate bothoutput 214 and aprobability distribution 222 associated withoutput 214 .

概率分布220-226可以分别与输出212-218相关。例如，一个或多个误差组件228可以使用与定位组件204相关联的一个或多个误差模型126来处理输出212，以便生成概率分布220。例如，如果输出212指出车辆102的位置，则概率分布220可以表示车辆102的估计位置，该估计位置是基于确定的位置和由针对定位组件204的一个或多个误差模型126表示的一个或多个误差的。此外，误差组件228可以使用与感知组件206相关联的一个或多个误差模型126来处理输出214，以便生成概率分布222。例如，如果输出214指出物体的速度，则概率分布222可以表示物体的可能速度，该可能速度是基于确定的速度和由针对感知组件206的一个或多个误差模型126表示的一个或多个误差的。Probability distributions 220-226 may be associated with outputs 212-218, respectively. For example, one ormore error components 228 may processoutput 212 using one ormore error models 126 associated withpositioning component 204 to generateprobability distribution 220 . For example, if theoutput 212 indicates the location of thevehicle 102 , theprobability distribution 220 may represent an estimated location of thevehicle 102 based on the determined location and one or more of the errors represented by the one ormore error models 126 for thelocation component 204 . an error. Additionally,error component 228 may processoutput 214 using one ormore error models 126 associated withperception component 206 to generateprobability distribution 222 . For example, ifoutput 214 indicates the speed of the object,probability distribution 222 may represent the likely speed of the object based on the determined speed and one or more errors represented by one ormore error models 126 forperception component 206 of.

估计组件230可以被配置为处理概率分布220-226中的一个或多个和/或传感器数据106(为了清楚起见并未示出)，以便生成与车辆102和/或物体相关联的估计位置232。如本文所讨论的，估计位置232可以包括位置的概率分布，例如，高斯分布。Estimation component 230 may be configured to process one or more of probability distributions 220 - 226 and/or sensor data 106 (not shown for clarity) in order to generate estimatedlocation 232 associated withvehicle 102 and/or object . As discussed herein, estimatedlocation 232 may include a probability distribution of locations, eg, a Gaussian distribution.

图3是根据本公开的实施例的车辆102使用一个或多个误差模型126分析传感器数据106以便确定与物体相关联的估计位置的另一示例的示图。在图3的示例中，估计组件230可以分析来自组件204-210中的一个或多个的输出212-218中的一个或多个，以便确定与车辆102和/或物体相关联的估计位置302。一个或多个误差组件228然后可以使用一个或多个误差模型126和估计位置302来确定车辆102和/或物体的估计位置304。3 is a diagram of another example of avehicle 102 analyzingsensor data 106 using one ormore error models 126 in order to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure. In the example of FIG. 3,estimation component 230 can analyze one or more of outputs 212-218 from one or more of components 204-210 in order to determine estimatedlocation 302 associated withvehicle 102 and/or object . The one ormore error components 228 may then use the one ormore error models 126 and the estimatedlocation 302 to determine the estimatedlocation 304 of thevehicle 102 and/or the object.

例如，一个或多个误差组件228可以使用一个或多个误差模型126来确定与一个或多个组件204-210的一个或多个输出212-218相关联的、用于确定估计位置302的一个或多个总误差和/或总误差百分比。一个或多个误差组件228然后可以使用一个或多个总误差和/或总误差百分比来生成估计位置304。如本文所讨论的，估计位置304可以包括位置的概率分布，例如，高斯分布。For example, one ormore error components 228 may use one ormore error models 126 to determine an associated one or more outputs 212-218 of one or more components 204-210 for use in determining estimatedposition 302 or multiple total errors and/or percentages of total errors. The one ormore error components 228 may then generate the estimatedposition 304 using the one or more total errors and/or the total error percentage. As discussed herein, estimatedlocation 304 may include a probability distribution of locations, eg, a Gaussian distribution.

图4是根据本公开的实施例的车辆102使用一个或多个不确定性模型130分析传感器数据106以便确定与物体相关联的估计位置的示例的示图。例如，一个或多个不确定性组件402可以被配置为使用一个或多个不确定性模型130来处理输出212-218，以便生成与输出212-218相关联的概率分布404-410。在一些情况下，一个或多个不确定性组件402可以包含于组件204-210内。例如，定位组件204可以分析传感器数据106，并且至少部分地基于该分析，生成输出212和与输出212相关联的概率分布404两者。对于另一个示例，感知组件206可以分析传感器数据106，并且至少部分地基于该分析，生成输出214和与输出214相关联的概率分布406两者。4 is an illustration of an example of avehicle 102 analyzingsensor data 106 using one ormore uncertainty models 130 in order to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure. For example, one ormore uncertainty components 402 may be configured to process outputs 212-218 using one ormore uncertainty models 130 to generate probability distributions 404-410 associated with outputs 212-218. In some cases, one ormore uncertainty components 402 may be included within components 204-210. For example, thepositioning component 204 can analyze thesensor data 106 and, based at least in part on the analysis, generate both theoutput 212 and theprobability distribution 404 associated with theoutput 212 . For another example,perception component 206 can analyzesensor data 106 and, based at least in part on the analysis, generate bothoutput 214 and aprobability distribution 406 associated withoutput 214 .

应注意的是，虽然在图4的示例中未示出，但是组件204-210中的一个或多个可以使用来自其他组件204-210中的一个或多个的输出212-218来生成输出212-218。例如，规划组件208可以使用来自定位组件204的输出212来生成输出216。对于另一个示例，规划组件208可以使用来自感知组件206的输出214来生成输出216。作为使用来自一个或多个其他组件204-210的输出212-218的补充或替代，组件204-210可以使用概率分布404-410。It should be noted that, although not shown in the example of FIG. 4, one or more of the components 204-210 may use the outputs 212-218 from one or more of the other components 204-210 to generate the output 212 -218. For example,planning component 208 can useoutput 212 frompositioning component 204 to generateoutput 216 . For another example,planning component 208 can useoutput 214 fromperception component 206 to generateoutput 216 . In addition to or instead of using the outputs 212-218 from one or more other components 204-210, the components 204-210 may use probability distributions 404-410.

概率分布404-410可以分别与输出212-218相关联。例如，一个或多个不确定性组件402可以使用与定位组件204相关联的一个或多个不确定性模型130来处理输出212，以便生成概率分布404。例如，如果输出212指出车辆102的位置，则概率分布404可以表示车辆102的估计位置，该估计位置是至少部分基于确定的位置和针对定位组件204的一个或多个不确定性模型130的。此外，一个或多个不确定性组件402可以使用与感知组件206相关联的一个或多个不确定性模型130来处理输出214，以便生成概率分布406。例如，如果输出214指出物体的速度，则概率分布406可以表示物体的可能速度，该可能速度是基于确定的速度和针对感知组件206的一个或多个不确定性模型130的。Probability distributions 404-410 may be associated with outputs 212-218, respectively. For example, one ormore uncertainty components 402 may processoutput 212 using one ormore uncertainty models 130 associated withpositioning component 204 to generateprobability distribution 404 . For example, ifoutput 212 indicates the location ofvehicle 102 ,probability distribution 404 may represent an estimated location ofvehicle 102 based at least in part on the determined location and one ormore uncertainty models 130 forpositioning component 204 . Additionally, one ormore uncertainty components 402 can processoutput 214 using one ormore uncertainty models 130 associated withperception component 206 to generateprobability distribution 406 . For example, ifoutput 214 indicates the speed of the object,probability distribution 406 may represent the likely speed of the object based on the determined speed and one ormore uncertainty models 130 forperception component 206 .

估计组件230可以被配置为处理概率分布404-410中的一个或多个和/或传感器数据106(为了清楚起见未示出)，以便生成与车辆102和/或物体相关联的估计位置412。如本文所讨论的，估计位置412可以包括位置的概率分布，例如，高斯分布。Estimation component 230 may be configured to process one or more of probability distributions 404 - 410 and/or sensor data 106 (not shown for clarity) to generate estimatedlocation 412 associated withvehicle 102 and/or object. As discussed herein, estimatedlocation 412 may include a probability distribution of locations, eg, a Gaussian distribution.

图5是根据本公开的实施例的车辆102使用一个或多个不确定性模型130分析传感器数据106以确定与物体相关联的估计位置的另一示例的示图。在图5的示例中，估计组件230可以分析来自组件204-210中的一个或多个的输出212-218中的一个或多个，以便确定与车辆102和/或物体相关联的估计位置302。一个或多个不确定性组件402然后可以使用一个或多个不确定性模型130和估计位置302来确定车辆102和/或物体的估计位置502。5 is a diagram of another example of avehicle 102 analyzingsensor data 106 using one ormore uncertainty models 130 to determine an estimated location associated with an object in accordance with an embodiment of the present disclosure. In the example of FIG. 5,estimation component 230 can analyze one or more of outputs 212-218 from one or more of components 204-210 in order to determine estimatedlocation 302 associated withvehicle 102 and/or object . The one ormore uncertainty components 402 may then use the one ormore uncertainty models 130 and the estimatedposition 302 to determine the estimatedposition 502 of thevehicle 102 and/or the object.

例如，一个或多个不确定性组件402可以使用组件204-210的一个或多个不确定性模型130来确定与一个或多个组件204-210的一个或多个输出212-218相关联的总不确定性，该总不确定性用于确定估计位置302。一个或多个不确定性组件402然后可以使用总不确定性来生成估计位置502。如本文所讨论的，估计位置502可以包括位置的概率分布，例如，高斯分布。For example, the one ormore uncertainty components 402 may use the one ormore uncertainty models 130 of the components 204-210 to determine the associated one or more outputs 212-218 of the one or more components 204-210 The total uncertainty used to determine the estimatedlocation 302 . The one ormore uncertainty components 402 can then use the total uncertainty to generate the estimatedposition 502 . As discussed herein, estimatedlocation 502 may include a probability distribution of locations, eg, a Gaussian distribution.

图6示出了根据本公开的实施例的说明车辆102确定一段时间内的碰撞概率的示例图表600。如图所示，图表600表示沿y轴的概率602和沿x轴的时间604。在图6的示例中，车辆102可以确定在时间606(1)的碰撞概率。例如，在时间606(1)，车辆102可以确定三个未来时间的碰撞概率，该三个未来时间为时间606(2)、时间606(3)和时间606(4)。在一些情况下，碰撞概率与车辆102和单个物体相关联。在其他情况下，碰撞概率与车辆102和多于一个物体相关联。FIG. 6 shows anexample graph 600 illustrating thevehicle 102 determining a probability of collision over a period of time in accordance with an embodiment of the present disclosure. As shown,graph 600 representsprobability 602 along the y-axis and time 604 along the x-axis. In the example of FIG. 6, thevehicle 102 may determine a probability of collision at time 606(1). For example, at time 606(1), thevehicle 102 may determine collision probabilities for three future times, time 606(2), time 606(3), and time 606(4). In some cases, the collision probability is associated with thevehicle 102 and a single object. In other cases, the collision probability is associated with thevehicle 102 and more than one object.

如图所示，车辆102可以确定在时间606(2)存在第一碰撞概率608(1)、在时间606(3)存在第二碰撞概率608(2)、并且在时间606(4)没有碰撞概率。第一碰撞概率608(1)可能与低风险相关联，第二碰撞概率608(2)可能与高风险相关联，并且由于在时间606(4)没有碰撞概率，所以在时间606(4)不存在碰撞风险。在一些情况下，至少部分地基于第一碰撞概率608(1)低于阈值概率，第一碰撞概率608(1)可以是低风险的。此外，至少部分地基于第二碰撞概率608(2)等于或大于阈值概率，第二碰撞概率608(2)可以是高风险的。As shown, thevehicle 102 may determine that there is a first probability of collision 608(1) at time 606(2), a second probability of collision 608(2) at time 606(3), and no collision at time 606(4) probability. The first collision probability 608(1) may be associated with a low risk, the second collision probability 608(2) may be associated with a high risk, and since there is no collision probability at time 606(4), there is no collision probability at time 606(4). There is a risk of collision. In some cases, the first collision probability 608(1) may be low risk based at least in part on the first collision probability 608(1) being below a threshold probability. Furthermore, the second collision probability 608(2) may be high risk based at least in part on the second collision probability 608(2) being equal to or greater than the threshold probability.

尽管图6的示例说明了确定在离散时间的碰撞概率，但在一些情况下，车辆102可以持续地确定碰撞概率。Although the example of FIG. 6 illustrates determining the probability of collision at discrete times, in some cases thevehicle 102 may continuously determine the probability of collision.

图7示出了根据本公开的实施例的至少部分地基于车辆数据和地面真值数据生成误差模型数据的示例700。如图7所描绘的，一个或多个车辆702可以生成车辆数据704并将车辆数据704传输到误差模型组件706。如本文所讨论的，误差模型组件706可以确定能够指示与参数相关联的误差的误差模型126。例如，车辆数据704可以是与一个或多个车辆702的组件相关联的数据，例如，感知组件206、规划组件208、定位组件204、估计组件230等。通过示例而非限制，车辆数据704可以与感知组件206相关联，并且车辆数据704可以包括与一个或多个车辆702在环境中检测到的物体相关联的边界框。7 illustrates an example 700 of generating error model data based at least in part on vehicle data and ground truth data, in accordance with an embodiment of the present disclosure. As depicted in FIG. 7 , one ormore vehicles 702 may generatevehicle data 704 and transmit thevehicle data 704 to anerror model component 706 . As discussed herein, theerror model component 706 can determine theerror model 126 that can indicate the error associated with the parameter. For example,vehicle data 704 may be data associated with one or more components ofvehicle 702, eg,perception component 206,planning component 208,localization component 204,estimation component 230, and the like. By way of example and not limitation,vehicle data 704 may be associated withperception component 206 andvehicle data 704 may include bounding boxes associated with one or more objects detected byvehicle 702 in the environment.

误差模型组件706可以接收可以手动标记和/或从其他经过验证的机器学习组件确定的地面真值数据708。通过示例而非限制，地面真值数据708可以包括与环境中的物体相关联的经过验证的边界框。通过将车辆数据704的边界框与地面真值数据708的边界框进行比较，误差模型组件706可以确定与一个或多个车辆702的系统(例如，组件)相关联的误差。这样的误差可以包括，例如，地面真值与输出之间的差异、百分比差异、误差率等。在一些情况下，车辆数据704可以包括与检测到的实体和/或该实体所在的环境相关联的一个或多个特性(也称为参数)。在一些示例中，与实体相关联的特性可以包括但不限于x位置(全局位置)、y位置(全局位置)、z位置(全局位置)、方位、实体类型(例如，分类)、实体的速率、实体的范围(大小)等。与环境相关联的特性可以包括但不限于环境中另一个实体的存在、环境中另一个实体的状态、一天中的某个时间、一周中的某一天、季节、天气状况、暗/明的指示等。因此，误差可能与其他特性(例如，环境参数)相关联。在至少一些示例中，可以为各种参数分组确定这样的误差模型(例如，针对分类、距离、速度等的不同组合的不同模型)。在至少一些示例中，这样的参数可以进一步包括环境信息，例如但不限于物体的数量、一天中的时间、一年中的时间、天气状况等。Theerror model component 706 can receiveground truth data 708 that can be manually labeled and/or determined from other validated machine learning components. By way of example and not limitation,ground truth data 708 may include validated bounding boxes associated with objects in the environment. By comparing the bounding boxes of thevehicle data 704 to the bounding boxes of theground truth data 708 , theerror model component 706 can determine errors associated with one or more systems (eg, components) of thevehicle 702 . Such errors may include, for example, differences between ground truth and output, percentage differences, error rates, and the like. In some cases,vehicle data 704 may include one or more characteristics (also referred to as parameters) associated with the detected entity and/or the environment in which the entity is located. In some examples, characteristics associated with an entity may include, but are not limited to, x position (global position), y position (global position), z position (global position), orientation, entity type (eg, classification), velocity of the entity , the extent (size) of the entity, etc. Properties associated with the environment may include, but are not limited to, the presence of another entity in the environment, the status of another entity in the environment, time of day, day of the week, season, weather conditions, indication of dark/light Wait. Therefore, errors may be associated with other characteristics (eg, environmental parameters). In at least some examples, such error models can be determined for various parameter groupings (eg, different models for different combinations of classification, distance, speed, etc.). In at least some examples, such parameters may further include environmental information such as, but not limited to, number of objects, time of day, time of year, weather conditions, and the like.

误差模型组件706可以处理多个车辆数据704和多个地面真值数据708以确定误差模型数据710。误差模型数据710可以包括由误差模型组件706计算的误差，其可以表示为误差712(1)-(3)。另外，误差模型组件706可以确定表示为概率714(1)-(3)的、与误差712(1)-(3)相关联的概率，其可以与用于呈现误差模型716(1)-(3)的环境参数相关联(其可以表示误差模型126)。通过示例而非限制，车辆数据704可以包括与在包括降雨的环境中距一个或多个车辆702在50米距离处的物体相关联的边界框。地面真值数据708可以提供与物体相关联的经验证的边界框。误差模型组件706可以确定误差模型数据710，该误差模型数据确定与一个或多个车辆702的感知系统相关联的误差。该50米距离和降雨可以作为环境参数来确定使用误差模型716(1)-(3)中的哪个误差模型。一旦识别出误差模型，误差模型716(1)-(3)可以基于概率714(1)-(3)提供误差712(1)-(3)，其中与较高概率714(1)-(3)相关联的误差712(1)-(3)相比与较低概率714(1)-(3)相关联的误差712(1)-(3)更有可能被选择。Theerror model component 706 can process the plurality ofvehicle data 704 and the plurality ofground truth data 708 to determineerror model data 710 .Error model data 710 may include errors calculated byerror model component 706, which may be represented as errors 712(1)-(3). Additionally,error model component 706 can determine probabilities associated with errors 712(1)-(3), represented as probabilities 714(1)-(3), which can be used in conjunction with presenting error models 716(1)-( 3) are associated with environmental parameters (which may represent the error model 126). By way of example and not limitation,vehicle data 704 may include bounding boxes associated with objects at a distance of 50 meters from one ormore vehicles 702 in an environment that includes rainfall.Ground truth data 708 may provide validated bounding boxes associated with objects. Theerror model component 706 can determineerror model data 710 that determines errors associated with the perception system of the one ormore vehicles 702 . The 50 meter distance and rainfall can be used as environmental parameters to determine which of the error models 716(1)-(3) to use. Once the error model is identified, the error model 716(1)-(3) may provide errors 712(1)-(3) based on the probabilities 714(1)-(3), where the same as the higher probability 714(1)-(3 ) associated errors 712(1)-(3) are more likely to be selected than errors 712(1)-(3) associated with lower probabilities 714(1)-(3).

图8示出了根据本公开的实施例的一个或多个车辆702生成车辆数据704并将车辆数据704传输到一个或多个计算设备802的示例800。如上所述，误差模型组件706可以确定可以指示与参数相关联的误差的感知误差模型。如上所述，车辆数据704可以包括由一个或多个车辆702的传感器生成的传感器数据和/或由一个或多个车辆702的感知系统生成的感知数据。可以通过将车辆数据704与地面真值数据708进行比较来确定感知误差模型。地面真值数据708可以被手动标记并且可以与环境相关，并且可以表示已知的结果。因此，车辆数据704中与地面真值数据708的偏差可以被识别为传感器系统和/或一个或多个车辆702的感知系统中的误差。通过示例而非限制，感知系统可以将物体识别为自行车骑士，其中地面真值数据708指示该物体为行人。通过另一示例而非限制，传感器系统可以生成将物体表示为具有2米宽度的传感器数据，其中地面真值数据708指示该物体具有3米宽度。8 illustrates an example 800 of one ormore vehicles 702 generatingvehicle data 704 and transmitting thevehicle data 704 to one or more computing devices 802 in accordance with embodiments of the present disclosure. As described above, theerror model component 706 can determine a perceptual error model that can be indicative of an error associated with a parameter. As discussed above,vehicle data 704 may include sensor data generated by one or more sensors ofvehicle 702 and/or perception data generated by one or more perception systems ofvehicle 702 . A perceptual error model may be determined by comparingvehicle data 704 to groundtruth data 708 . Theground truth data 708 can be manually labeled and correlated with the environment and can represent known outcomes. Accordingly, deviations invehicle data 704 fromground truth data 708 may be identified as errors in sensor systems and/or perception systems of one ormore vehicles 702 . By way of example and not limitation, the perception system may identify an object as a cyclist, where theground truth data 708 indicates that the object is a pedestrian. By way of another example and not limitation, a sensor system may generate sensor data representing an object as having a width of 2 meters, wherein theground truth data 708 indicates that the object has a width of 3 meters.

如上所述，误差模型组件706可以确定与车辆数据704中表示的物体相关联的分类，并且确定车辆数据704和/或其他记录数据中具有相同分类的其他物体。然后误差模型组件706可以确定与误差范围相关联的概率分布，该误差范围与物体相关联。基于比较和误差范围，误差模型组件706可以确定估计位置502。As described above, theerror model component 706 can determine the classification associated with the object represented in thevehicle data 704 and determine other objects in thevehicle data 704 and/or other recorded data having the same classification. Theerror model component 706 can then determine a probability distribution associated with the error range associated with the object. Based on the comparison and the error range, theerror model component 706 can determine the estimatedlocation 502 .

如图8中描绘的，环境804可以包括表示为由感知系统生成的边界框的物体806(1)-(3)。感知误差模型数据808可以将情境参数指示为810(1)-(3)并且将与情境参数相关联的误差指示为812(1)-(3)。As depicted in FIG. 8,environment 804 may include objects 806(1)-(3) represented as bounding boxes generated by the perception system. Perceptual error model data 808 may indicate context parameters as 810(1)-(3) and errors associated with the contextual parameters as 812(1)-(3).

图9示出了根据本公开的实施例的至少部分地基于车辆数据和地面真值数据生成不确定性数据的示例900。如图9中描绘的，一个或多个车辆702可以生成车辆数据704并将车辆数据704传输到不确定性模型组件902。如本文所讨论的，不确定性模型组件902可以确定与确定参数的组件相关联的不确定性。例如，车辆数据704可以是与一个或多个车辆702的组件相关联的数据，该组件例如为感知组件206、规划组件208、定位组件204、预测组件210等。通过示例而非限制，车辆数据704可以与感知组件206相关联，并且车辆数据704可以包括与由一个或多个车辆702在环境中检测到的物体相关联的边界框。9 illustrates an example 900 of generating uncertainty data based at least in part on vehicle data and ground truth data, in accordance with an embodiment of the present disclosure. As depicted in FIG. 9 , one ormore vehicles 702 may generatevehicle data 704 and transmit thevehicle data 704 to anuncertainty model component 902 . As discussed herein, theuncertainty model component 902 can determine the uncertainty associated with the component that determines the parameter. For example,vehicle data 704 may be data associated with one or more components ofvehicle 702, such asperception component 206,planning component 208,localization component 204,prediction component 210, and the like. By way of example and not limitation,vehicle data 704 may be associated withperception component 206 andvehicle data 704 may include bounding boxes associated with objects detected in the environment by one ormore vehicles 702 .

不确定性模型组件902可以接收可以被手动标记和/或从其他经过验证的机器学习组件确定的地面真值数据708。通过示例而非限制，地面真值数据708可以包括与环境中的物体相关联的经过验证的边界框。通过比较车辆数据704与地面真值数据708，不确定性模型组件902可以确定一个或多个车辆702的系统(例如，组件)确定地面真值的一致性。例如，该一致性可以指示由车辆数据704表示的参数的百分比与由地面真值数据708表示的参数的百分比相同。Uncertainty model component 902 can receiveground truth data 708 that can be manually labeled and/or determined from other validated machine learning components. By way of example and not limitation,ground truth data 708 may include validated bounding boxes associated with objects in the environment. By comparing thevehicle data 704 to theground truth data 708, theuncertainty model component 902 can determine the consistency of one or more systems (eg, components) of thevehicle 702 in determining the ground truth. For example, the agreement may indicate that the percentage of parameters represented byvehicle data 704 is the same as the percentage of parameters represented byground truth data 708 .

不确定性模型组件902然后可以使用该一致性来生成与确定参数的组件相关联和/或与确定参数的组件相关联的不确定性数据904。例如，如果该一致性指示存在低百分比，则不确定性数据904可以指示高不确定性。然而，如果一致性数据指示存在高百分比，则不确定性数据904可以指示低不确定性。Theuncertainty model component 902 can then use the agreement to generateuncertainty data 904 associated with and/or associated with the parameter-determining component. For example, if the agreement indicates that there is a low percentage, then theuncertainty data 904 may indicate a high uncertainty. However, if the agreement data indicates that there is a high percentage, then theuncertainty data 904 may indicate a low uncertainty.

更详细地，不确定性模型组件902可以识别一种或多种类型的不确定性。不确定性的类型可以包括但不限于认知不确定性、偶然不确定性(例如，数据相关、任务相关等)等。认知不确定性可能与对组件生成的数据的无知相关联。偶然不确定性可能与关于数据无法解释的信息的不确定性相关联。不确定性模型组件902然后可以使用识别的一个或多个不确定性来生成一个或多个不确定性模型130。In more detail, theuncertainty model component 902 can identify one or more types of uncertainty. Types of uncertainty may include, but are not limited to, epistemic uncertainty, contingent uncertainty (eg, data-related, task-related, etc.), and the like. Cognitive uncertainty can be associated with ignorance of data generated by components. Accidental uncertainty can be associated with uncertainty about information that the data cannot explain. Theuncertainty model component 902 can then use the identified one or more uncertainties to generate one ormore uncertainty models 130 .

在一些情况下，不确定性模型组件902可以多次地将数据输入到组件中，其中组件的一个或多个节点在输入数据时发生变化，这导致组件的输出不同。这可能会导致来自组件的输出中的范围。在一些情况下，组件可以进一步将输出的均值和/或方差输出。不确定性模型组件902然后可以使用与输出的范围、均值和/或方差相关联的分布，以针对组件和/或输出类型(例如，参数)生成一个或多个不确定性模型130。In some cases, theuncertainty model component 902 may input data into the component multiple times, where one or more nodes of the component change when the data is input, which results in different outputs of the component. This can cause ranges in the output from the component. In some cases, the component may further output the mean and/or variance of the output. Theuncertainty model component 902 can then use the distributions associated with the ranges, means, and/or variances of the outputs to generate one ormore uncertainty models 130 for components and/or output types (eg, parameters).

图10描绘了用于实现本文所讨论的技术的示例系统1000的方框图。在至少一个示例中，系统1000可以包括车辆102。在所示示例1000中，车辆102为自动驾驶车辆；然而，车辆102可以是任何其他类型的车辆(例如，可以提供执行各种操控是否安全的指示的、由驾驶员控制的车辆)。10 depicts a block diagram of anexample system 1000 for implementing the techniques discussed herein. In at least one example,system 1000 may includevehicle 102 . In the illustrated example 1000, thevehicle 102 is an autonomous vehicle; however, thevehicle 102 may be any other type of vehicle (eg, a driver-controlled vehicle that may provide indications of whether it is safe to perform various maneuvers).

车辆102可以包括一个或多个计算设备1002、一个或多个传感器系统202、一个或多个发射器1004、一个或多个通信连接1006(也称为通信设备和/或调制解调器)、至少一个直接连接1008(例如，用于与车辆102物理耦合以交换数据和/或提供电力)、以及一个或多个驱动系统1010。一个或多个传感器系统202可以被配置为获取与环境相关联的传感器数据106。Thevehicle 102 may include one ormore computing devices 1002, one ormore sensor systems 202, one or more transmitters 1004, one or more communication connections 1006 (also referred to as communication devices and/or modems), at least one direct Connections 1008 (eg, for physically coupling with thevehicle 102 to exchange data and/or provide power), and one or more drive systems 1010 . One ormore sensor systems 202 may be configured to acquiresensor data 106 associated with the environment.

一个或多个传感器系统202可以包括飞行时间传感器、位置传感器(例如，GPS、指南针等)、惯性传感器(例如，惯性测量单元(IMU)、加速度计、磁力计、陀螺仪等)、激光雷达传感器、雷达传感器、声纳传感器、红外传感器、摄像机(例如，RGB、IR、强度、深度等)、麦克风传感器、环境传感器(例如，温度传感器、湿度传感器、光传感器、压力传感器等)、超声波转换器、车轮编码器等。一个或多个传感器系统202可以包括这些或其他类型的传感器中的每一个的多个实例。例如，飞行时间传感器可以包括位于车辆102的角部、前部、后部、侧面和/或顶部的单独的飞行时间传感器。作为另一示例，摄像机传感器可以包括设置于车辆102的外部和/或内部周围的不同位置的多个摄像机。一个或多个传感器系统202可以向一个或多个计算设备1002提供输入。One ormore sensor systems 202 may include time-of-flight sensors, location sensors (eg, GPS, compass, etc.), inertial sensors (eg, inertial measurement units (IMUs), accelerometers, magnetometers, gyroscopes, etc.), lidar sensors , radar sensors, sonar sensors, infrared sensors, cameras (eg, RGB, IR, intensity, depth, etc.), microphone sensors, environmental sensors (eg, temperature sensors, humidity sensors, light sensors, pressure sensors, etc.), ultrasonic transducers , wheel encoder, etc. One ormore sensor systems 202 may include multiple instances of each of these or other types of sensors. For example, the time-of-flight sensors may include separate time-of-flight sensors located at the corners, front, rear, sides, and/or roof of thevehicle 102 . As another example, the camera sensor may include multiple cameras disposed at various locations around the exterior and/or interior of thevehicle 102 . One ormore sensor systems 202 may provide input to one ormore computing devices 1002 .

车辆102还可以包括一个或多个发射器1004，其用于发射光和/或声音。该示例中的一个或多个发射器1004包括用于与车辆102的乘客通信的内部音频和视觉发射器。通过示例而非限制，内部发射器可以包括扬声器、灯、标志、显示屏、触摸屏、触觉发射器(例如，振动和/或力反馈)、机械式致动器(例如，安全带张紧器、座椅定位器、头枕定位器等)等。该示例中的一个或多个发射器1004还包括外部发射器。通过示例而非限制，在该示例中，外部发射器包括用信号通知行进方向的灯或车辆动作的其他指示器(例如，指示器灯、标志、灯组等)，以及一个或多个音频发射器(例如，扬声器、扬声器组、喇叭等)，以与行人或其他附近的车辆进行声音通信，其中一项或多项可以包括声束转向技术。Thevehicle 102 may also include one or more transmitters 1004 for emitting light and/or sound. The one or more transmitters 1004 in this example include interior audio and visual transmitters for communicating with occupants of thevehicle 102 . By way of example and not limitation, internal transmitters may include speakers, lights, signs, display screens, touch screens, haptic transmitters (eg, vibration and/or force feedback), mechanical actuators (eg, seat belt tensioners, seat positioner, headrest positioner, etc.), etc. The one or more transmitters 1004 in this example also include external transmitters. By way of example and not limitation, in this example, external transmitters include lights that signal direction of travel or other indicators of vehicle action (eg, indicator lights, signs, groups of lights, etc.), and one or more audio emissions devices (eg, speakers, speaker groups, horns, etc.) for acoustic communication with pedestrians or other nearby vehicles, one or more of which may include beam steering technology.

车辆102还可以包括一个或多个通信连接1006，其实现车辆102与一个或多个其他本地或远程计算设备(例如，远程长距离操作计算设备)或远程服务之间的通信。例如，一个或多个通信连接1006可以促进与车辆102和/或一个或多个驱动系统1010上的一个或多个其他本地计算设备的通信。此外，一个或多个通信连接1006可以允许车辆102与其他附近的一个或多个计算设备(例如，其他附近的车辆、交通信号等)进行通信。Thevehicle 102 may also include one or more communication connections 1006 that enable communication between thevehicle 102 and one or more other local or remote computing devices (eg, remote long-range operating computing devices) or remote services. For example, one or more communication connections 1006 may facilitate communication with one or more other local computing devices onvehicle 102 and/or one or more drive systems 1010 . Additionally, one or more communication connections 1006 may allow thevehicle 102 to communicate with one or more other nearby computing devices (eg, other nearby vehicles, traffic signals, etc.).

一个或多个通信连接1006可以包括用于将一个或多个计算设备1002连接到另一计算设备或一个或多个外部网络1012(例如，因特网)的物理和/或逻辑接口。例如，一个或多个通信连接1006可以实现基于Wi-Fi的通信，例如，经由IEEE802.11标准定义的频率、短程无线频率(例如，蓝牙)、移动通信(例如，2G、3G、4G、4GLTE、5G等)、卫星通信、专用短程通信(DSRC)、或任何合适的有线或无线通信协议，其使得相应的计算设备能够与一个或多个其他计算设备相接合。在至少一些示例中，一个或多个通信连接1006可以包括如上面详细描述的一个或多个调制解调器。One or more communication connections 1006 may include physical and/or logical interfaces for connecting one ormore computing devices 1002 to another computing device or to one or more external networks 1012 (eg, the Internet). For example, one or more communication connections 1006 may enable Wi-Fi based communications, eg, via frequencies defined by the IEEE 802.11 standard, short-range wireless frequencies (eg, Bluetooth), mobile communications (eg, 2G, 3G, 4G, 4G LTE) , 5G, etc.), satellite communication, dedicated short-range communication (DSRC), or any suitable wired or wireless communication protocol that enables a corresponding computing device to interface with one or more other computing devices. In at least some examples, one or more communication connections 1006 may include one or more modems as described in detail above.

在至少一个示例中，车辆102可以包括一个或多个驱动系统1010。在一些示例中，车辆102可以具有单一驱动系统1010。在至少一个示例中，如果车辆102具有多个驱动系统1010，则各个驱动系统1010可以定位于车辆102的相对端(例如，前部和后部等)。在至少一个示例中，一个或多个驱动系统1010可以包括一个或多个传感器系统202以检测一个或多个驱动系统1010和/或车辆102周围环境的状况。通过示例而非限制，一个或多个传感器系统202可以包括用于感测驱动系统的车轮的旋转的一个或多个车轮编码器(例如，旋转编码器)，用于测量驱动系统的方位和加速度的惯性传感器(例如，惯性测量单元、加速度计、陀螺仪、磁力计等)，摄像机或其他图像传感器，用于声学地检测驱动系统周围环境中的物体的超声波传感器，激光雷达传感器，雷达传感器等。一些传感器，例如，车轮编码器，可以是一个或多个驱动系统1010唯一的。在一些情况下，一个或多个驱动系统1010上的一个或多个传感器系统202可以与车辆102的对应系统(例如，一个或多个传感器系统202)重叠或对其进行补充。In at least one example, thevehicle 102 may include one or more drive systems 1010 . In some examples,vehicle 102 may have a single drive system 1010 . In at least one example, if thevehicle 102 has multiple drive systems 1010 , each drive system 1010 may be positioned at opposite ends of the vehicle 102 (eg, front and rear, etc.). In at least one example, one or more drive systems 1010 may include one ormore sensor systems 202 to detect conditions of one or more drive systems 1010 and/or the environment surrounding thevehicle 102 . By way of example and not limitation, the one ormore sensor systems 202 may include one or more wheel encoders (eg, rotary encoders) for sensing rotation of the wheels of the drive system, for measuring the orientation and acceleration of the drive system inertial sensors (e.g. inertial measurement units, accelerometers, gyroscopes, magnetometers, etc.), cameras or other image sensors, ultrasonic sensors for acoustically detecting objects in the environment surrounding the drive system, lidar sensors, radar sensors, etc. . Some sensors, such as wheel encoders, may be unique to one or more drive systems 1010 . In some cases, one ormore sensor systems 202 on one or more drive systems 1010 may overlap or complement corresponding systems (eg, one or more sensor systems 202 ) ofvehicle 102 .

一个或多个驱动系统1010可以包括许多车辆系统，包括高压电池、驱动车辆的电机、将来自电池的直流电转换为交流电以供其他车辆系统使用的变流器、包括转向电机和转向齿条(其可以是电动的)的转向系统、包括液压或电动致动器的制动系统、包括液压和/或气动组件的悬挂系统、用于分配制动力以减轻抓地力损失并保持控制的稳定控制系统、HVAC系统、照明(例如，诸如用于照亮车辆外部周围环境的头灯/尾灯之类的照明)以及一个或多个其他系统(例如，冷却系统、安全系统、车载充电系统、其他电气组件，例如，DC/DC转换器、高压接头、高压电缆、充电系统、充电端口等)。另外，一个或多个驱动系统1010可以包括驱动系统控制器，其可以接收和预处理来自一个或多个传感器系统202的数据并且控制各种车辆系统的操作。在一些示例中，驱动系统控制器可以包括一个或多个处理器和与一个或多个处理器通信地耦合的存储器。存储器可以存储一个或多个模块以执行一个或多个驱动系统1010的各种功能。此外，一个或多个驱动系统1010还包括一个或多个通信连接，其使得相应的驱动系统能够与一个或多个其他本地或远程计算设备进行通信。One or more drive systems 1010 may include a number of vehicle systems, including a high voltage battery, an electric motor that drives the vehicle, an inverter that converts DC power from the battery to AC power for use by other vehicle systems, including a steering motor and a steering rack (which Steering systems that may be electric), braking systems including hydraulic or electric actuators, suspension systems including hydraulic and/or pneumatic components, stability control systems for distributing braking force to mitigate grip loss and maintain control, HVAC systems, lighting (eg, lighting such as headlights/taillights used to illuminate the surrounding environment outside the vehicle), and one or more other systems (eg, cooling systems, safety systems, on-board charging systems, other electrical components, For example, DC/DC converters, high voltage connectors, high voltage cables, charging systems, charging ports, etc.). Additionally, one or more drive systems 1010 may include a drive system controller that may receive and preprocess data from one ormore sensor systems 202 and control the operation of various vehicle systems. In some examples, the drive system controller may include one or more processors and a memory communicatively coupled with the one or more processors. The memory may store one or more modules to perform various functions of one or more drive system 1010 . Additionally, one or more drive systems 1010 also include one or more communication connections that enable the respective drive systems to communicate with one or more other local or remote computing devices.

一个或多个计算设备1002可以包括一个或多个处理器1014和与处理器1014通信地耦合的存储器1016。在所示示例中，一个或多个计算设备1002的存储器1016可以存储定位组件204、感知组件206、预测组件210、估计组件230、规划组件208、一个或多个误差组件228、一个或多个不确定性组件402和一个或多个传感器系统202。尽管出于说明目的描述为驻留于存储器1016中，但是可以设想，定位组件204、感知组件206、预测组件210、估计组件230、规划组件208、一个或多个误差组件228、一个或多个不确定性组件402和一个或多个系统控制器1018可以另外地或可替代地可由一个或多个计算设备1002访问(例如，存储在车辆102的不同组件中和/或可由车辆102访问(例如，远程存储)。One ormore computing devices 1002 may include one ormore processors 1014 and amemory 1016 communicatively coupled with theprocessors 1014 . In the example shown,memory 1016 of one ormore computing devices 1002 may storepositioning component 204,sensing component 206, predictingcomponent 210, estimatingcomponent 230,planning component 208, one ormore error components 228, one ormore Uncertainty component 402 and one ormore sensor systems 202 . Although described as residing inmemory 1016 for illustrative purposes, it is contemplated thatlocation component 204,perception component 206,prediction component 210,estimation component 230,planning component 208, one ormore error components 228, one or more Theuncertainty component 402 and the one or more system controllers 1018 may additionally or alternatively be accessible by the one or more computing devices 1002 (eg, stored in various components of thevehicle 102 and/or accessible by the vehicle 102 (eg, , remote storage).

在一个或多个计算设备1002的存储器1016中，定位组件204可以包括从一个或多个传感器系统202接收数据以确定车辆102的位置的功能。例如，定位组件204可以包括和/或请求/接收环境的三维地图，并且可以持续确定自动驾驶车辆在地图内的位置。在一些情况下，定位组件204可以使用SLAM(同时定位和映射)或CLAMS(同时校准、定位和映射)来接收飞行时间数据、图像数据、激光雷达数据、雷达数据、声纳数据、IMU数据、GPS数据、车轮编码器数据或其任意组合等，以准确地确定自动驾驶车辆的位置。在一些情况下，定位组件204可以向车辆102的各个组件提供数据以确定自动驾驶车辆的初始位置以用于生成轨迹，如本文中所讨论的。In thememory 1016 of the one ormore computing devices 1002 , thelocation component 204 may include functionality to receive data from the one ormore sensor systems 202 to determine the location of thevehicle 102 . For example, thepositioning component 204 can include and/or request/receive a three-dimensional map of the environment, and can continuously determine the location of the autonomous vehicle within the map. In some cases, thelocalization component 204 may use SLAM (simultaneous localization and mapping) or CLAMS (simultaneous calibration, localization and mapping) to receive time-of-flight data, image data, lidar data, radar data, sonar data, IMU data, GPS data, wheel encoder data, or any combination thereof, etc., to accurately determine the location of an autonomous vehicle. In some cases, thepositioning component 204 may provide data to various components of thevehicle 102 to determine the initial position of the autonomous vehicle for use in generating the trajectory, as discussed herein.

感知组件206可以包括执行物体检测、分割和/或分类的功能。在一些示例中，感知组件206可以提供经处理的传感器数据，该数据指示存在靠近车辆102的实体和/或将实体分类为实体类型(例如，汽车、行人、自行车骑士、建筑物、树木、路面、路缘、人行道、未知等)。在附加和/或替代示例中，感知组件206可以提供经处理的传感器数据，该数据指示与检测到的实体和/或该实体所处的环境相关联的一个或多个特性(也称为参数)。在一些示例中，与实体相关联的特性可以包括但不限于x位置(全局位置)、y位置(全局位置)、z位置(全局位置)、方位、实体类型(例如，分类)、实体的速率、实体的范围(大小)等。与环境相关联的特性可以包括但不限于环境中存在另一个实体、环境中另一个实体的状态、一天中的某个时间、一周中的一天、季节、天气状况、地理位置、暗/明的指示等。Perception component 206 may include functionality to perform object detection, segmentation, and/or classification. In some examples,perception component 206 can provide processed sensor data that indicates the presence of entities proximate tovehicle 102 and/or classifies entities as entity types (eg, car, pedestrian, cyclist, building, tree, road surface) , curb, sidewalk, unknown, etc.). In additional and/or alternative examples,perception component 206 may provide processed sensor data indicative of one or more characteristics (also referred to as parameters) associated with the detected entity and/or the environment in which the entity is located ). In some examples, characteristics associated with an entity may include, but are not limited to, x position (global position), y position (global position), z position (global position), orientation, entity type (eg, classification), velocity of the entity , the extent (size) of the entity, etc. Properties associated with the environment may include, but are not limited to, the presence of another entity in the environment, the state of another entity in the environment, time of day, day of the week, season, weather conditions, geographic location, dark/light instructions etc.

感知组件206可以包括存储由感知组件206生成的感知数据的功能。在一些情况下，感知组件206可以确定对应于已分类为物体类型的物体的追踪。仅出于说明的目的，使用一个或多个传感器系统202，感知组件206可以获取环境的一个或多个图像。一个或多个传感器系统202可以获取包括诸如行人之类的物体的环境的图像。行人可以在时间T处于第一位置并且在时间T+t处于第二位置(例如，在时间T之后的时间t跨度内的移动)。换言之，行人可以在这段时间跨度内从第一位置移动到第二位置。例如，这样的运动可以记录为与物体相关联的存储的感知数据。Theperception component 206 may include functionality to store perception data generated by theperception component 206 . In some cases,perception component 206 can determine a track corresponding to an object that has been classified as an object type. For purposes of illustration only, using one ormore sensor systems 202,perception component 206 may acquire one or more images of the environment. One ormore sensor systems 202 may acquire images of an environment including objects such as pedestrians. The pedestrian may be in a first position at time T and a second position at time T+t (eg, movement within a time span of time t after time T). In other words, the pedestrian can move from the first position to the second position within this time span. For example, such motion may be recorded as stored sensory data associated with the object.

在一些示例中，存储的感知数据可以包括由车辆获取的融合感知数据。融合感知数据可以包括来自一个或多个传感器系统202的传感器数据的融合或其他组合，该传感器系统例如为图像传感器、激光雷达传感器、雷达传感器、飞行时间传感器、声纳传感器、全球定位系统传感器、内部传感器和/或这些传感器的任意组合。存储的感知数据可以另外地或可替代地包括分类数据，其包括传感器数据中表示的物体(例如，行人、车辆、建筑物、路面等)的语义分类。存储的感知数据可以另外地或可替代地包括与分类为动态物体的物体在环境中的运动相对应的追踪数据(随时间与物体相关联的历史位置、方位、传感器特征等的集合)。追踪数据可以包括多个不同物体随时间的追踪。当物体静止(例如，静止不动)或移动(例如，步行、跑步等)时，可以挖掘该追踪数据以识别某些类型的物体(例如，行人、动物等)的图像。在该示例中，计算设备确定对应于行人的追踪。In some examples, the stored sensory data may include fused sensory data acquired by the vehicle. Fused perception data may include fusion or other combinations of sensor data from one ormore sensor systems 202, such as image sensors, lidar sensors, radar sensors, time-of-flight sensors, sonar sensors, global positioning system sensors, Internal sensors and/or any combination of these sensors. The stored sensory data may additionally or alternatively include classification data, which includes semantic classifications of objects (eg, pedestrians, vehicles, buildings, road surfaces, etc.) represented in the sensor data. Stored sensory data may additionally or alternatively include tracking data (a collection of historical positions, orientations, sensor characteristics, etc. associated with objects over time) corresponding to the movement of objects classified as dynamic objects in the environment. Tracking data may include tracking of a number of different objects over time. This tracking data can be mined to identify images of certain types of objects (eg, pedestrians, animals, etc.) when the object is stationary (eg, stationary) or moving (eg, walking, running, etc.). In this example, the computing device determines the tracking corresponding to the pedestrian.

预测组件210可以生成表示环境中的一个或多个物体的估计位置的预测概率的一个或多个概率图。例如，预测组件210可以针对距车辆102在阈值距离内的车辆、行人、动物等生成一个或多个概率图。在一些情况下，预测组件210可以测量物体的追踪并基于观察和预测的行为生成物体的离散化预测概率图、热学图、概率分布、离散化概率分布和/或轨迹。在一些情况下，一个或多个概率图可以表示环境中一个或多个物体的意图。Theprediction component 210 can generate one or more probability maps representing predicted probabilities of estimated locations of one or more objects in the environment. For example, theprediction component 210 can generate one or more probability maps for vehicles, pedestrians, animals, etc. within a threshold distance from thevehicle 102 . In some cases, theprediction component 210 can measure the tracking of the object and generate a discretized predicted probability map, thermal map, probability distribution, discretized probability distribution, and/or trajectory of the object based on observed and predicted behavior. In some cases, the one or more probability maps may represent the intent of one or more objects in the environment.

规划组件208可以确定车辆102要依循以穿越环境的路径。例如，规划组件208可以确定各种路线和路径以及各种细节水平。在一些情况下，规划组件208可以确定从第一位置(例如，当前位置)行进到第二位置(例如，目标位置)的路线。出于讨论的目的，路线可以是在两个位置之间行进的航路点的序列。作为非限制性示例，航路点包括街道、交叉路口、全球定位系统(GPS)的坐标等。此外，规划组件208可以生成用于沿着从第一位置到第二位置的路线的至少一部分引导车辆102的指令。在至少一个示例中，规划组件208可以确定如何将车辆102从航路点的序列中的第一航路点引导至航路点的序列中的第二航路点。在一些示例中，指令可以是路径或路径的一部分。在一些示例中，可以根据滚动域技术大致上同时生成多条路径(即，在技术容差内)。可以选择具有最高置信水平的滚动数据域内的多条路径中的一条路径来操作车辆。Theplanning component 208 can determine a path to be followed by thevehicle 102 to traverse the environment. For example,planning component 208 can determine various routes and paths and various levels of detail. In some cases,planning component 208 can determine a route to travel from a first location (eg, a current location) to a second location (eg, a target location). For discussion purposes, a route may be a sequence of waypoints traveled between two locations. By way of non-limiting example, waypoints include streets, intersections, global positioning system (GPS) coordinates, and the like. Additionally, theplanning component 208 can generate instructions for guiding thevehicle 102 along at least a portion of the route from the first location to the second location. In at least one example, theplanning component 208 can determine how to steer thevehicle 102 from a first waypoint in the sequence of waypoints to a second waypoint in the sequence of waypoints. In some examples, the instruction may be a path or part of a path. In some examples, multiple paths may be generated substantially simultaneously (ie, within technical tolerances) according to rolling domain techniques. One of the paths within the rolling data domain with the highest confidence level may be selected to operate the vehicle.

在其他示例中，规划组件208可以可替代地或另外地，使用来自感知组件206和/或预测组件210的数据来确定车辆102要依循以穿越环境的路径。例如，规划组件208和/或预测组件210可以从感知组件206接收关于与环境相关联的物体的数据。使用该数据，规划组件208可以确定从第一位置(例如，当前位置)行进到第二位置(例如，目标位置)的路线，以避开环境中的物体。在至少一些示例中，规划组件208可以确定不存在这样的无碰撞路径，并且进而可以提供将车辆102带入安全停止的路径，从而避免所有碰撞和/或以其他方式减轻损坏。In other examples,planning component 208 may alternatively or additionally use data from sensingcomponent 206 and/or predictingcomponent 210 to determine a path to be followed byvehicle 102 to traverse the environment. For example,planning component 208 and/orprediction component 210 may receive data fromperception component 206 regarding objects associated with the environment. Using this data,planning component 208 can determine a route to travel from a first location (eg, a current location) to a second location (eg, a target location) to avoid objects in the environment. In at least some examples,planning component 208 can determine that no such collision-free path exists, and in turn can provide a path to bringvehicle 102 to a safe stop, thereby avoiding all collisions and/or otherwise mitigating damage.

在至少一个示例中，一个或多个计算设备1002可以包括一个或多个系统控制器1018，其可以被配置为控制车辆102的转向、推进、制动、安全、发射器、通信和其他系统。这些系统控制器1018可以与一个或多个驱动系统1010的对应系统和/或车辆102的其他组件通信和/或对其进行控制，其可以被配置为根据从规划组件208提供的路径进行操作。In at least one example, one ormore computing devices 1002 may include one or more system controllers 1018 , which may be configured to control steering, propulsion, braking, safety, transmitter, communications, and other systems ofvehicle 102 . These system controllers 1018 may communicate with and/or control one or more corresponding systems of the drive system 1010 and/or other components of thevehicle 102 , which may be configured to operate according to paths provided from theplanning component 208 .

车辆102可以通过一个或多个网络1012连接到一个或多个计算设备802并且可以包括一个或多个处理器1020和与一个或多个处理器820通信地耦合的存储器1022。在至少一种情况下，一个或多个处理器820可以类似于一个或多个处理器1014，并且存储器1022可以类似于存储器1016。在所示示例中，一个或多个计算设备802的存储器1022可以存储车辆数据704、地面真值数据708和误差模型组件706。尽管出于说明的目的被描述为驻留于存储器1022中，但是可以设想，车辆数据704、地面真值数据708和/或误差模型组件706可以另外地或可替代地可由一个或多个计算设备802访问(例如，存储在一个或多个计算设备802的不同组件中和/或可由一个或多个计算设备802访问(例如，远程存储))。Thevehicle 102 may be connected to one or more computing devices 802 through one or more networks 1012 and may include one or more processors 1020 and a memory 1022 communicatively coupled with the one or more processors 820 . In at least one instance, the one or more processors 820 can be similar to the one ormore processors 1014 , and the memory 1022 can be similar to thememory 1016 . In the example shown, memory 1022 of one or more computing devices 802 may storevehicle data 704 ,ground truth data 708 , anderror model component 706 . Although described as residing in memory 1022 for illustrative purposes, it is contemplated thatvehicle data 704,ground truth data 708, and/orerror model component 706 may additionally or alternatively be accessible by one or more computing devices 802 (eg, stored in and/or accessible by one or more computing devices 802 (eg, stored remotely) in various components of one or more computing devices 802 ).

一个或多个计算设备1002的一个或多个处理器1014和一个或多个计算设备802的一个或多个处理器1020可以是能够执行指令以处理数据并执行如本文所述的操作的任何合适的处理器。通过示例而非限制，一个或多个处理器1014和1020可以包括一个或多个中央处理单元(CPU)、图形处理单元(GPU)，或处理电子数据以将该电子数据转换成可以存储在寄存器和/或存储器中的其他电子数据的任何其他设备或设备的一部分。在一些示例中，集成电路(例如，ASIC等)、门阵列(例如，FPGA等)和其他硬件设备只要被配置为实现编码指令，也可以被认为是处理器。The one ormore processors 1014 of the one ormore computing devices 1002 and the one or more processors 1020 of the one or more computing devices 802 may be any suitable capable of executing instructions to process data and perform operations as described herein processor. By way of example and not limitation, one ormore processors 1014 and 1020 may include one or more central processing units (CPUs), graphics processing units (GPUs), or process electronic data to convert the electronic data into registers that can be stored in registers and/or other electronic data in memory on any other device or part of the device. In some examples, integrated circuits (eg, ASICs, etc.), gate arrays (eg, FPGAs, etc.), and other hardware devices may also be considered processors as long as they are configured to implement encoded instructions.

一个或多个计算设备1002的存储器1016和一个或多个计算设备802的存储器1022是非暂时性计算机可读介质的示例。存储器1016和1022可以存储操作系统和一个或多个软件应用、指令、程序和/或数据以实现本文描述的方法和属于各种系统的功能。在各种实施方式中，存储器1016和1022可以使用任何合适的存储器技术来实现，例如，静态随机存取存储器(SRAM)、同步动态RAM(SDRAM)、非易失性/闪存型存储器或能够存储信息的任何其他类型的存储器。本文描述的架构、系统和单独的元素可以包括许多其他逻辑、程序和物理组件，其中附图中所示的那些仅仅是与此处的讨论相关联的示例。Thememory 1016 of the one ormore computing devices 1002 and the memory 1022 of the one or more computing devices 802 are examples of non-transitory computer-readable media.Memories 1016 and 1022 may store an operating system and one or more software applications, instructions, programs and/or data to implement the methods and functions pertaining to various systems described herein. In various implementations,memories 1016 and 1022 may be implemented using any suitable memory technology, such as static random access memory (SRAM), synchronous dynamic RAM (SDRAM), non-volatile/flash-type memory, or capable of storing any other type of storage of information. The architectures, systems, and individual elements described herein may include many other logical, procedural, and physical components, of which those shown in the figures are merely examples associated with the discussions herein.

在一些情况下，本文讨论的一些或所有组件的方面可以包括任何模型、算法和/或机器学习算法。例如，在一些情况下，存储器1016和1022中的组件可以实现为神经网络。In some cases, aspects of some or all of the components discussed herein may include any model, algorithm, and/or machine learning algorithm. For example, in some cases, components inmemories 1016 and 1022 may be implemented as neural networks.

图11-14示出了根据本公开的实施例的示例过程。这些过程被示出为逻辑流程图，其中的每个操作表示可以以硬件、软件或其组合实现的操作序列。在软件的上下文中，操作表示存储于一个或多个计算机可读存储介质上的计算机可执行指令，该指令当由一个或多个处理器执行时，执行所阐述的操作。一般地，计算机可执行指令包括执行特定功能或实现特定抽像数据类型的例程、程序、对象、组件、数据结构等。描述操作的顺序不旨在被解释为限制，并且任何数量的描述的操作可以以任何顺序和/或并行组合以实现该过程。11-14 illustrate example processes in accordance with embodiments of the present disclosure. The processes are shown as logic flow diagrams, where each operation represents a sequence of operations that can be implemented in hardware, software, or a combination thereof. In the context of software, operations refer to computer-executable instructions stored on one or more computer-readable storage media that, when executed by one or more processors, perform the recited operations. Generally, computer-executable instructions include routines, programs, objects, components, data structures, etc. that perform particular functions or implement particular abstract data types. The order in which the operations are described is not intended to be construed as limiting, and any number of the described operations may be combined in any order and/or in parallel to implement the process.

图11描绘了根据本公开的实施例的用于使用误差模型执行碰撞监测的示例过程1100。在操作1102处，过程1100可以包括接收由一个或多个传感器生成的传感器数据。例如，车辆102可以沿着从第一位置到第二位置的路径导航。在导航时，车辆102可以使用车辆102的一个或多个传感器生成传感器数据。FIG. 11 depicts anexample process 1100 for performing collision monitoring using an error model, according to an embodiment of the present disclosure. At operation 1102,process 1100 may include receiving sensor data generated by one or more sensors. For example, thevehicle 102 may navigate along a path from a first location to a second location. While navigating, thevehicle 102 may generate sensor data using one or more sensors of thevehicle 102 .

在操作1104处，过程1100可以包括至少使用车辆的第一系统，至少部分地基于传感器数据的第一部分来至少确定与车辆相关联的参数。例如，车辆102可以使用一个或多个系统分析传感器数据的第一部分。一个或多个系统可以包括但不限于定位系统、感知系统、规划系统、预测系统等。至少部分地基于分析，车辆102可以确定与车辆102相关联的参数。参数可以包括但不限于车辆102的位置、车辆102的速度、车辆102的行进方向等。Atoperation 1104 ,process 1100 may include using at least a first system of the vehicle to determine at least a parameter associated with the vehicle based at least in part on the first portion of the sensor data. For example, thevehicle 102 may analyze the first portion of the sensor data using one or more systems. The one or more systems may include, but are not limited to, positioning systems, perception systems, planning systems, prediction systems, and the like. Based at least in part on the analysis, thevehicle 102 may determine parameters associated with thevehicle 102 . Parameters may include, but are not limited to, the position of thevehicle 102, the speed of thevehicle 102, the direction of travel of thevehicle 102, and the like.

在操作1106处，过程1100可以包括至少部分地基于与车辆相关联的参数和与第一系统相关联的第一误差模型来确定与车辆相关联的估计位置。例如，车辆102可以使用第一误差模型至少处理与车辆102相关联的参数。如本文所讨论的，第一误差模型可以表示与第一系统的输出相关联的误差和/或误差百分比。至少部分地基于该处理，车辆102可以确定在之后时间的与车辆102相关联的估计位置。如本文还讨论的，估计位置可以对应于位置的概率分布。Atoperation 1106 , theprocess 1100 may include determining an estimated position associated with the vehicle based at least in part on parameters associated with the vehicle and a first error model associated with the first system. For example, thevehicle 102 may process at least parameters associated with thevehicle 102 using the first error model. As discussed herein, the first error model may represent an error and/or a percentage of error associated with the output of the first system. Based at least in part on this process, thevehicle 102 may determine an estimated location associated with thevehicle 102 at a later time. As also discussed herein, the estimated location may correspond to a probability distribution of locations.

在操作1108处，过程1100可以包括至少使用车辆的第二系统，至少部分地基于传感器数据的第二部分来至少确定与物体相关联的参数。例如，车辆102可以分析传感器数据并且至少部分地基于该分析来识别物体。车辆102然后可以使用一个或多个系统来分析传感器数据的第二部分。至少部分地基于分析，车辆102可以确定与物体相关联的参数。参数可以包括但不限于物体的类型、物体的位置、物体的速度、物体的行进方向等。Atoperation 1108 ,process 1100 may include using at least a second system of the vehicle to determine at least a parameter associated with the object based at least in part on the second portion of the sensor data. For example, thevehicle 102 may analyze sensor data and identify objects based at least in part on the analysis. Thevehicle 102 may then analyze the second portion of the sensor data using one or more systems. Based at least in part on the analysis, thevehicle 102 may determine parameters associated with the object. Parameters may include, but are not limited to, the type of object, the location of the object, the speed of the object, the direction of travel of the object, and the like.

在操作1110处，过程1100可以包括至少部分地基于与物体相关联的参数和与第二系统相关联的第二误差模型来确定与物体相关联的估计位置。例如，车辆102可以使用第二误差模型至少处理与物体相关联的参数。如本文所讨论的，第二误差模型可以表示与第二系统的输出相关联的误差和/或误差百分比。至少部分地基于该处理，车辆102可以确定在之后时间的与物体相关联的估计位置。如本文还讨论的，估计位置可以对应于位置的概率分布。Atoperation 1110, theprocess 1100 may include determining an estimated location associated with the object based at least in part on parameters associated with the object and a second error model associated with the second system. For example, thevehicle 102 may use the second error model to process at least the parameters associated with the object. As discussed herein, the second error model may represent an error and/or a percentage of error associated with the output of the second system. Based at least in part on this process, thevehicle 102 may determine an estimated location associated with the object at a later time. As also discussed herein, the estimated location may correspond to a probability distribution of locations.

在操作1112处，过程1100可以包括至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置来确定碰撞概率。例如，车辆102可以分析与车辆102相关联的估计位置和与物体相关联的估计位置以便确定碰撞概率。在一些情况下，碰撞概率可以至少部分地基于与车辆102相关联的估计位置和与物体相关联的估计位置之间的重叠量。Atoperation 1112, theprocess 1100 may include determining a probability of collision based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object. For example, thevehicle 102 may analyze the estimated location associated with thevehicle 102 and the estimated location associated with the object in order to determine a probability of collision. In some cases, the probability of collision may be based, at least in part, on an amount of overlap between the estimated location associated with thevehicle 102 and the estimated location associated with the object.

在操作1114处，过程1100可以包括确定碰撞概率是否等于或大于阈值。例如，车辆102可以将碰撞概率与阈值进行比较以便确定碰撞概率是否等于或大于阈值。Atoperation 1114, theprocess 1100 may include determining whether a collision probability is equal to or greater than a threshold. For example, thevehicle 102 may compare the collision probability to a threshold to determine whether the collision probability is equal to or greater than the threshold.

如果在操作1114处确定碰撞概率不等于或大于阈值，则在操作1116处，过程1100可以包括使车辆继续沿路径导航。例如，如果车辆102确定碰撞概率小于阈值，则车辆102可以继续沿路径导航。If it is determined atoperation 1114 that the probability of collision is not equal to or greater than the threshold, atoperation 1116 theprocess 1100 may include continuing the vehicle to navigate along the path. For example, if thevehicle 102 determines that the probability of collision is less than a threshold, thevehicle 102 may continue to navigate the path.

然而，如果在操作1114处确定碰撞概率等于或大于阈值，则在操作1118处，过程1100可以包括使车辆执行一个或多个动作。例如，如果车辆102确定碰撞概率等于或大于阈值，则车辆102可以执行一个或多个动作。该一个或多个动作可以包括但不限于改变车辆102的路径、改变车辆102的速度、使车辆102停车等。However, if it is determined atoperation 1114 that the probability of collision is equal to or greater than the threshold, atoperation 1118 theprocess 1100 may include causing the vehicle to perform one or more actions. For example, if thevehicle 102 determines that the collision probability is equal to or greater than a threshold, thevehicle 102 may perform one or more actions. The one or more actions may include, but are not limited to, changing the path of thevehicle 102, changing the speed of thevehicle 102, parking thevehicle 102, and the like.

图12描绘了根据本公开的实施例的用于使用误差模型来确定与物体相关联的估计位置的示例过程1200。在操作1202处，过程1200可以包括接收由一个或多个传感器生成的传感器数据。例如，车辆102可以沿着从第一位置到第二位置的路径导航。在导航时，车辆102可以使用车辆102的一个或多个传感器生成传感器数据。12 depicts anexample process 1200 for using an error model to determine an estimated location associated with an object, according to an embodiment of the present disclosure. Atoperation 1202,process 1200 may include receiving sensor data generated by one or more sensors. For example, thevehicle 102 may navigate along a path from a first location to a second location. While navigating, thevehicle 102 may generate sensor data using one or more sensors of thevehicle 102 .

在操作1204处，过程1200可以包括使用车辆的一个或多个系统，至少部分地基于传感器数据来确定与物体相关联的第一参数。例如，车辆102可以使用一个或多个系统来分析传感器数据。该一个或多个系统可以包括但不限于，定位系统、感知系统、规划系统、预测系统等。至少部分地基于该分析，车辆102可以确定与物体(例如，车辆或另一物体)相关联的第一参数。第一参数可以包括但不限于物体的位置、物体的速度、物体的行进方向等。Atoperation 1204 ,process 1200 may include using one or more systems of the vehicle to determine a first parameter associated with the object based at least in part on the sensor data. For example, thevehicle 102 may use one or more systems to analyze sensor data. The one or more systems may include, but are not limited to, positioning systems, perception systems, planning systems, prediction systems, and the like. Based at least in part on the analysis, thevehicle 102 may determine a first parameter associated with the object (eg, the vehicle or another object). The first parameter may include, but is not limited to, the position of the object, the speed of the object, the direction of travel of the object, and the like.

在操作1206处，过程1200可以包括至少部分地基于第一误差模型来确定与第一参数相关联的第一概率分布。例如，车辆102可以使用第一误差模型处理至少第一参数。如本文所讨论的，第一误差模型可以表示与第一参数相关联的误差和/或误差百分比。至少部分地基于该处理，车辆102可以确定与第一参数相关联的第一概率分布。Atoperation 1206, theprocess 1200 may include determining a first probability distribution associated with the first parameter based at least in part on the first error model. For example, thevehicle 102 may process at least the first parameter using the first error model. As discussed herein, the first error model may represent an error and/or a percentage of error associated with the first parameter. Based at least in part on the process, thevehicle 102 may determine a first probability distribution associated with the first parameter.

在操作1208处，过程1200可以包括使用车辆的一个或多个系统，至少部分地基于传感器数据或第一概率分布中的至少一个来确定与物体相关联的第二参数。例如，车辆102可以使用一个或多个系统来分析传感器数据或第一概率分布中的至少一个。在一些情况下，当第二参数是使用第一参数确定的时，车辆102对第一概率分布进行分析。至少部分地基于该分析，车辆102可以确定与物体(例如，车辆或另一物体)相关联的第二参数。第二参数可以包括但不限于物体的位置、物体的速度、物体的行进方向、物体在未来时间的估计位置等。Atoperation 1208 ,process 1200 may include using one or more systems of the vehicle to determine a second parameter associated with the object based at least in part on at least one of the sensor data or the first probability distribution. For example, thevehicle 102 may use one or more systems to analyze at least one of the sensor data or the first probability distribution. In some cases, thevehicle 102 analyzes the first probability distribution when the second parameter is determined using the first parameter. Based at least in part on the analysis, thevehicle 102 may determine a second parameter associated with the object (eg, the vehicle or another object). The second parameter may include, but is not limited to, the position of the object, the speed of the object, the direction of travel of the object, the estimated position of the object at a future time, and the like.

在操作1210处，过程1200可以包括至少部分地基于第二误差模型来确定与第二参数相关联的第二概率分布。例如，车辆102可以使用第二误差模型至少处理第二参数。如本文所讨论的，第二误差模型可以表示与第二参数相关联的误差和/或误差百分比。至少部分地基于该处理，车辆102可以确定与第二参数相关联的第二概率分布。Atoperation 1210, theprocess 1200 can include determining a second probability distribution associated with the second parameter based at least in part on the second error model. For example, thevehicle 102 may process at least the second parameter using the second error model. As discussed herein, the second error model may represent an error and/or a percentage of error associated with the second parameter. Based at least in part on the process, thevehicle 102 may determine a second probability distribution associated with the second parameter.

在操作1212处，过程1200可以包括至少部分地基于第一概率分布或第二概率分布中的至少一个来确定与物体相关联的估计位置。例如，车辆102可以至少部分地基于第一概率分布和/或第二概率分布来确定估计位置。在一些情况下，如果第一参数和第二参数是独立的，例如，第一参数指示物体的当前位置而第二参数指示物体的速度，则车辆102可以使用第一概率分布和第二概率分布两者来确定估计位置。在一些情况下，如果第二参数是使用第一参数确定的，例如，如果第二参数指示物体在未来时间的估计位置，而该估计位置是使用指示物体的速度的第一参数确定的，则车辆102可以使用第二概率分布来确定估计位置。Atoperation 1212, theprocess 1200 may include determining an estimated location associated with the object based at least in part on at least one of the first probability distribution or the second probability distribution. For example, thevehicle 102 may determine the estimated location based at least in part on the first probability distribution and/or the second probability distribution. In some cases, thevehicle 102 may use the first and second probability distributions if the first and second parameters are independent, eg, the first parameter indicates the current position of the object and the second parameter indicates the speed of the object both to determine the estimated location. In some cases, if the second parameter is determined using the first parameter, eg, if the second parameter indicates an estimated position of the object at a future time, and the estimated position is determined using the first parameter indicating the velocity of the object, then Thevehicle 102 may use the second probability distribution to determine the estimated location.

图13A-13B描绘了根据本公开的实施例的用于使用不确定性模型执行碰撞监测的示例过程1300。在操作1302处，过程1300可以包括接收由一个或多个传感器生成的传感器数据。例如，车辆102可以沿着从第一位置到第二位置的路径导航。在导航时，车辆102可以使用车辆102的一个或多个传感器生成传感器数据。13A-13B depict anexample process 1300 for performing collision monitoring using an uncertainty model, according to an embodiment of the present disclosure. At operation 1302,process 1300 may include receiving sensor data generated by one or more sensors. For example, thevehicle 102 may navigate along a path from a first location to a second location. While navigating, thevehicle 102 may generate sensor data using one or more sensors of thevehicle 102 .

在操作1304处，过程1300可以包括至少使用车辆的第一系统，至少部分地基于传感器数据的第一部分来至少确定与车辆相关联的参数。例如，车辆102可以使用一个或多个系统来分析传感器数据的第一部分。一个或多个系统可以包括但不限于定位系统、感知系统、规划系统、预测系统等。至少部分地基于该分析，车辆102可以确定与车辆102相关联的参数。该参数可以包括但不限于车辆102的位置、车辆102的速度、车辆102的行进方向等。Atoperation 1304 ,process 1300 may include determining at least a parameter associated with the vehicle based at least in part on the first portion of the sensor data, using at least a first system of the vehicle. For example, thevehicle 102 may analyze the first portion of the sensor data using one or more systems. The one or more systems may include, but are not limited to, positioning systems, perception systems, planning systems, prediction systems, and the like. Based at least in part on the analysis, thevehicle 102 may determine parameters associated with thevehicle 102 . The parameters may include, but are not limited to, the position of thevehicle 102, the speed of thevehicle 102, the direction of travel of thevehicle 102, and the like.

在操作1306处，过程1300可以包括确定与第一系统确定与车辆相关联的参数相关联的第一不确定性模型。例如，车辆102可以确定第一不确定性模型。在一些情况下，车辆102通过从第一系统接收第一不确定性模型来确定第一不确定性模型。在一些情况下，车辆102使用不确定性数据确定第一不确定性模型，不确定性数据指示与第一系统确定第一参数相关联的不确定性。Atoperation 1306 , theprocess 1300 may include determining a first uncertainty model associated with the first system-determined parameter associated with the vehicle. For example, thevehicle 102 may determine a first uncertainty model. In some cases, thevehicle 102 determines the first uncertainty model by receiving the first uncertainty model from the first system. In some cases, thevehicle 102 determines the first uncertainty model using uncertainty data indicative of uncertainty associated with the first system's determination of the first parameter.

在操作1308处，过程1300可以包括至少部分地基于与车辆相关联的参数和第一不确定性模型来确定与车辆相关联的估计位置。例如，车辆102可以使用第一不确定性模型至少处理与车辆102相关联的参数。至少部分地基于该处理，车辆102可以确定在之后时间的与车辆102相关联的估计位置。还如本文所讨论的，估计位置可以对应于位置的概率分布。Atoperation 1308 , theprocess 1300 may include determining an estimated location associated with the vehicle based at least in part on parameters associated with the vehicle and the first uncertainty model. For example, thevehicle 102 may process at least parameters associated with thevehicle 102 using the first uncertainty model. Based at least in part on this process, thevehicle 102 may determine an estimated location associated with thevehicle 102 at a later time. As also discussed herein, the estimated location may correspond to a probability distribution of locations.

在操作1310处，过程1300可以包括至少使用车辆的第二系统，至少部分地基于传感器数据的第二部分来至少确定与物体相关联的参数。例如，车辆102可以分析传感器数据并且至少部分地基于该分析来识别物体。车辆102然后可以使用一个或多个系统来分析传感器数据的第二部分。至少部分地基于该分析，车辆102可以确定与物体相关联的参数。参数可以包括但不限于物体的类型、物体的位置、物体的速度、物体的行进方向等。Atoperation 1310,process 1300 may include determining at least a parameter associated with the object based at least in part on the second portion of the sensor data using at least a second system of the vehicle. For example, thevehicle 102 may analyze sensor data and identify objects based at least in part on the analysis. Thevehicle 102 may then analyze the second portion of the sensor data using one or more systems. Based at least in part on the analysis, thevehicle 102 may determine parameters associated with the object. Parameters may include, but are not limited to, the type of object, the location of the object, the speed of the object, the direction of travel of the object, and the like.

在操作1312处，过程1300可以包括确定与第二系统确定与物体相关联的参数相关联的第二不确定性模型。例如，车辆102可以确定第二不确定性模型。在一些情况下，车辆102通过从第二系统接收第二不确定性模型来确定第二不确定性模型。在一些情况下，车辆102使用不确定性数据确定第二不确定性模型，不确定性数据指示与第二系统确定第二参数相关联的不确定性。Atoperation 1312, theprocess 1300 may include determining a second uncertainty model associated with the second system determining the parameter associated with the object. For example, thevehicle 102 may determine a second uncertainty model. In some cases, thevehicle 102 determines the second uncertainty model by receiving the second uncertainty model from the second system. In some cases, thevehicle 102 determines the second uncertainty model using the uncertainty data indicating the uncertainty associated with the second system's determination of the second parameter.

在操作1314处，过程1300可以包括至少部分地基于与物体相关联的参数和第二不确定性模型来确定与物体相关联的估计位置。例如，车辆102可以使用第二不确定性模型至少处理与物体相关联的参数。至少部分地基于该处理，车辆102可以确定在之后时间的与物体相关联的估计位置。还如本文所讨论的，估计位置可以对应于位置的概率分布。Atoperation 1314,process 1300 may include determining an estimated location associated with the object based at least in part on parameters associated with the object and a second uncertainty model. For example, thevehicle 102 may process at least parameters associated with the object using the second uncertainty model. Based at least in part on this process, thevehicle 102 may determine an estimated location associated with the object at a later time. As also discussed herein, the estimated location may correspond to a probability distribution of locations.

在操作1316处，过程1300可以包括至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置来确定碰撞概率。例如，车辆102可以分析与车辆102相关联的估计位置和与物体相关联的估计位置以便确定碰撞概率。在一些情况下，碰撞概率可以至少部分地基于与车辆102相关联的估计位置和与物体相关联的估计位置之间的重叠量。Atoperation 1316, theprocess 1300 may include determining a probability of collision based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object. For example, thevehicle 102 may analyze the estimated location associated with thevehicle 102 and the estimated location associated with the object in order to determine a probability of collision. In some cases, the probability of collision may be based, at least in part, on an amount of overlap between the estimated location associated with thevehicle 102 and the estimated location associated with the object.

在操作1318处，过程1300可以包括确定碰撞概率是否等于或大于阈值。例如，车辆102可以将碰撞概率与阈值进行比较以确定碰撞概率是否等于或大于阈值。Atoperation 1318,process 1300 may include determining whether a collision probability is equal to or greater than a threshold. For example, thevehicle 102 may compare the collision probability to a threshold to determine whether the collision probability is equal to or greater than the threshold.

如果在操作1318处确定碰撞概率不等于或大于阈值，则在操作1320处，过程1300可以包括使车辆继续沿着路径导航。例如，如果车辆102确定碰撞概率小于阈值，则车辆102可以继续沿着路径导航。If it is determined atoperation 1318 that the probability of collision is not equal to or greater than the threshold, atoperation 1320 theprocess 1300 may include continuing the vehicle to navigate along the path. For example, if thevehicle 102 determines that the collision probability is less than a threshold, thevehicle 102 may continue to navigate along the path.

然而，如果在操作1318处确定碰撞概率等于或大于阈值，则在操作1322处，过程1300可以包括使车辆执行一个或多个动作。例如，如果车辆102确定碰撞概率等于或大于阈值，则车辆102可以执行一个或多个动作。该一个或多个动作可以包括但不限于改变车辆102的路径、改变车辆102的速度、使车辆102停车等。However, if atoperation 1318 it is determined that the probability of collision is equal to or greater than the threshold, atoperation 1322 theprocess 1300 may include causing the vehicle to perform one or more actions. For example, if thevehicle 102 determines that the collision probability is equal to or greater than a threshold, thevehicle 102 may perform one or more actions. The one or more actions may include, but are not limited to, changing the path of thevehicle 102, changing the speed of thevehicle 102, parking thevehicle 102, and the like.

需要注意的是，在一些示例中，车辆102可以使用与车辆102相关联的多条可能路线来执行步骤1304-1314。在这样的示例中，车辆102可以选择包括最低不确定性和/或最低碰撞概率的路线。Note that in some examples, thevehicle 102 may perform steps 1304 - 1314 using multiple possible routes associated with thevehicle 102 . In such an example, thevehicle 102 may select a route that includes the lowest uncertainty and/or lowest probability of collision.

图14描绘了根据本公开的实施例的用于使用不确定性模型来确定与物体相关联的估计位置的示例过程1400。在操作1402处，过程1400可以包括接收由一个或多个传感器生成的传感器数据。例如，车辆102可以沿着从第一位置到第二位置的路径导航。在导航时，车辆102可以使用车辆102的一个或多个传感器生成传感器数据。14 depicts anexample process 1400 for determining an estimated location associated with an object using an uncertainty model, according to an embodiment of the present disclosure. Atoperation 1402,process 1400 may include receiving sensor data generated by one or more sensors. For example, thevehicle 102 may navigate along a path from a first location to a second location. While navigating, thevehicle 102 may generate sensor data using one or more sensors of thevehicle 102 .

在操作1404处，过程1400可以包括使用车辆的一个或多个系统，至少部分地基于传感器数据来确定与物体相关联的第一参数。例如，车辆102可以使用一个或多个系统来分析传感器数据。一个或多个系统可以包括但不限于定位系统、感知系统、规划系统、预测系统等。至少部分地基于该分析，车辆102可以确定与物体(例如，车辆或另一物体)相关联的第一参数。第一参数可以包括但不限于物体的位置、物体的速度、物体的行进方向等。Atoperation 1404 ,process 1400 may include using one or more systems of the vehicle to determine a first parameter associated with the object based at least in part on the sensor data. For example, thevehicle 102 may use one or more systems to analyze sensor data. The one or more systems may include, but are not limited to, positioning systems, perception systems, planning systems, prediction systems, and the like. Based at least in part on the analysis, thevehicle 102 may determine a first parameter associated with the object (eg, the vehicle or another object). The first parameter may include, but is not limited to, the position of the object, the speed of the object, the direction of travel of the object, and the like.

在操作1406处，过程1400可以包括至少部分地基于第一不确定性模型来确定与第一参数相关联的第一概率分布。例如，车辆102可以使用第一不确定性模型至少处理第一参数。至少部分地基于该处理，车辆102可以确定与第一参数相关联的第一概率分布。Atoperation 1406, theprocess 1400 may include determining a first probability distribution associated with the first parameter based at least in part on the first uncertainty model. For example, thevehicle 102 may process at least the first parameter using the first uncertainty model. Based at least in part on the process, thevehicle 102 may determine a first probability distribution associated with the first parameter.

在操作1408处，过程1400可以包括使用车辆的一个或多个系统，至少部分地基于传感器数据或第一概率分布中的至少一个来确定与物体相关联的第二参数。例如，车辆102可以使用一个或多个系统来分析传感器数据或第一概率分布中的至少一个。在一些情况下，当第二参数是使用第一参数确定的时，车辆102可以对第一概率分布进行分析。至少部分地基于该分析，车辆102可以确定与物体(例如，车辆或另一物体)相关联的第二参数。第二参数可以包括但不限于物体的位置、物体的速度、物体的行进方向、物体在未来时间的估计位置等。Atoperation 1408 ,process 1400 may include using one or more systems of the vehicle to determine a second parameter associated with the object based at least in part on at least one of the sensor data or the first probability distribution. For example, thevehicle 102 may use one or more systems to analyze at least one of the sensor data or the first probability distribution. In some cases, thevehicle 102 may analyze the first probability distribution when the second parameter is determined using the first parameter. Based at least in part on the analysis, thevehicle 102 may determine a second parameter associated with the object (eg, the vehicle or another object). The second parameter may include, but is not limited to, the position of the object, the speed of the object, the direction of travel of the object, the estimated position of the object at a future time, and the like.

在操作1410处，过程1400可以包括至少部分地基于第二不确定性模型来确定与第二参数相关联的第二概率分布。例如，车辆102可以使用第二不确定性模型至少处理第二参数。至少部分地基于该处理，车辆102可以确定与第二参数相关联的第二概率分布。Atoperation 1410, theprocess 1400 may include determining a second probability distribution associated with the second parameter based at least in part on the second uncertainty model. For example, thevehicle 102 may process at least the second parameter using the second uncertainty model. Based at least in part on the process, thevehicle 102 may determine a second probability distribution associated with the second parameter.

在操作1412处，过程1400可以包括至少部分地基于第一概率分布或第二概率分布中的至少一个来确定与物体相关联的估计位置。例如，车辆102可以至少部分地基于第一概率分布和/或第二概率分布来确定估计位置。在一些情况下，如果第一参数和第二参数是独立的，例如，第一参数指示物体的当前位置而第二参数指示物体的速度，则车辆102可以使用第一概率分布和第二概率分布两者来确定估计位置。在一些情况下，如果第二参数是使用第一参数来确定的，例如，如果第二参数指示物体在未来时间的估计位置，而该估计位置是使用指示物体的速度的第一参数确定的，则车辆102可以使用第二概率分布来确定估计位置。Atoperation 1412, theprocess 1400 may include determining an estimated location associated with the object based at least in part on at least one of the first probability distribution or the second probability distribution. For example, thevehicle 102 may determine the estimated location based at least in part on the first probability distribution and/or the second probability distribution. In some cases, thevehicle 102 may use the first and second probability distributions if the first and second parameters are independent, eg, the first parameter indicates the current position of the object and the second parameter indicates the speed of the object both to determine the estimated location. In some cases, if the second parameter is determined using the first parameter, for example, if the second parameter indicates an estimated position of the object at a future time, and the estimated position is determined using the first parameter indicating the velocity of the object, Thevehicle 102 may then use the second probability distribution to determine the estimated location.

结论in conclusion

虽然已描述了本文描述的技术的一个或多个示例，但其各种变化、增加、排列和等效对象仍包含于本文描述的技术的范围内。While one or more examples of the techniques described herein have been described, various modifications, additions, permutations, and equivalents thereof are intended to be included within the scope of the techniques described herein.

在示例的描述中，参考构成其一部分的附图，附图通过说明的方式示出了要求保护的主题的具体示例。应当理解的是，可以使用其他示例并且可以对其进行变更或替换，例如，结构上的变更。这样的示例、变更或替换不一定会偏离关于预期要求保护的主题的范围。虽然本文的步骤可能会按特定顺序呈现，但在一些情况下，顺序可能会改变，因此在不改变所描述的系统和方法的功能的情况下，可以在不同时间或以不同顺序提供某些输入。所公开的程序也能以不同的顺序执行。此外，本文中的各种计算不一定按照公开的顺序执行，并且可以容易地实现使用计算的替代排序的其他示例。除了重排序之外，计算还可以分解为具有相同结果的子计算。In the description of the examples, reference is made to the accompanying drawings, which form a part hereof, showing by way of illustration specific examples of the claimed subject matter. It is to be understood that other examples may be utilized and that changes or substitutions may be made, eg, structural changes. Such examples, changes or substitutions do not necessarily depart from the scope with respect to the intended claimed subject matter. Although the steps herein may be presented in a particular order, in some cases the order may be altered so that certain inputs may be provided at different times or in different orders without altering the functionality of the described systems and methods . The disclosed procedures can also be executed in a different order. Furthermore, the various calculations herein are not necessarily performed in the order disclosed, and other examples using alternative orderings of the calculations may be readily implemented. In addition to reordering, computations can be decomposed into sub-computations with the same result.

示例条款Sample Clause

A：一种自动驾驶车辆，包括：一个或多个传感器；一个或多个处理器；以及一种或多种计算机可读介质，其存储指令，指令当由一个或多个处理器执行时，使得一个或多个处理器执行操作，操作包括：从一个或多个传感器获取传感器数据；至少部分地基于传感器数据的第一部分，确定自动驾驶车辆在未来时间的估计位置；至少部分地基于自动驾驶车辆的系统和传感器数据的第二部分，确定物体在未来时间的估计位置；至少部分地基于误差模型和物体的估计位置，确定与物体相关联的估计位置的分布，误差模型表示与系统相关联的误差概率；至少部分地基于自动驾驶车辆的估计位置以及与物体相关联的估计位置的分布，确定自动驾驶车辆和物体之间的碰撞概率；并且至少部分地基于碰撞概率，使得自动驾驶车辆执行一个或多个动作。A: An autonomous vehicle comprising: one or more sensors; one or more processors; and one or more computer-readable media storing instructions that, when executed by the one or more processors, causing the one or more processors to perform operations comprising: obtaining sensor data from the one or more sensors; determining an estimated location of the autonomous vehicle at a future time based at least in part on the first portion of the sensor data; based at least in part on the autonomous driving a second portion of system and sensor data for the vehicle, determining an estimated position of the object at a future time; based at least in part on an error model and the estimated position of the object, determining a distribution of estimated positions associated with the object, the error model representing the associated with the system determining a probability of collision between the autonomous vehicle and the object based at least in part on the estimated position of the autonomous vehicle and the distribution of the estimated positions associated with the object; and based at least in part on the probability of collision, causing the autonomous vehicle to perform one or more actions.

B：如段落A中引述的自动驾驶车辆，操作还包括从一个或多个计算设备接收误差模型，误差模型是至少使用由一个或多个车辆生成的传感器数据来生成的。B: The autonomous vehicle as recited in paragraph A, the operations further comprising receiving an error model from one or more computing devices, the error model being generated using at least sensor data generated by the one or more vehicles.

C：如段落A或B中引述的自动驾驶车辆，操作还包括：至少部分地基于附加误差模型和车辆的估计位置，确定与自动驾驶车辆相关联的估计位置的分布；并且其中，确定自动驾驶车辆和物体之间的碰撞概率至少包括：确定与自动驾驶车辆相关联的估计位置的分布和与物体相关联的估计位置的分布之间的重叠量；并且至少部分地基于重叠量，确定碰撞概率。C: The autonomous vehicle as recited in paragraph A or B, the operations further comprising: determining a distribution of estimated positions associated with the autonomous vehicle based at least in part on the additional error model and the estimated positions of the vehicle; and wherein determining the autonomous driving The probability of collision between the vehicle and the object includes at least: determining an amount of overlap between a distribution of estimated locations associated with the autonomous vehicle and a distribution of estimated locations associated with the object; and based at least in part on the amount of overlap, determining a probability of collision .

D：如段落A至C中任一项引述的自动驾驶车辆，其中：至少部分地基于自动驾驶车辆的附加系统，进一步确定物体在未来时间的估计位置；并且至少部分地基于附加误差模型，进一步确定估计位置的分布，附加误差模型表示与附加系统相关联的误差分布。D: The autonomous vehicle as recited in any of paragraphs A to C, wherein: an estimated location of the object at a future time is further determined based at least in part on an additional system of the autonomous vehicle; and based at least in part on an additional error model, further A distribution of estimated locations is determined, and the additional error model represents the distribution of errors associated with the additional system.

E：一种方法，包括：从车辆的一个或多个传感器接收传感器数据；至少部分地基于传感器数据的第一部分，确定在一时间的与车辆相关联的估计位置；至少部分地基于车辆的系统和传感器数据的第二部分，确定与物体相关联的参数；至少部分地基于误差模型和与物体相关联的参数，确定在时间的与物体相关联的估计位置，误差模型表示与系统相关联的误差概率；并且至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置，使得车辆执行一个或多个动作。E: A method comprising: receiving sensor data from one or more sensors of a vehicle; determining, based at least in part on a first portion of the sensor data, an estimated location associated with the vehicle at a time; based at least in part on a system of the vehicle and a second portion of the sensor data, determining a parameter associated with the object; determining an estimated location associated with the object at time based at least in part on an error model and the parameter associated with the object, the error model representing a parameter associated with the system an error probability; and causing the vehicle to perform one or more actions based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object.

F：如段落E中引述的方法，还包括从一个或多个计算设备接收误差模型，误差模型是至少使用由一个或多个车辆生成的传感器数据来生成的。F: The method as recited in paragraph E, further comprising receiving an error model from one or more computing devices, the error model being generated using at least sensor data generated by the one or more vehicles.

G：如段落E或F中引述的方法，其中，参数包括以下各项中的至少一项：与物体相关联的物体类型；物体在环境内的位置；物体的速度；或物体在环境内的行进方向。G: A method as recited in paragraph E or F, wherein the parameters include at least one of: an object type associated with the object; the object's position within the environment; the object's velocity; or the object's speed within the environment direction of travel.

H：如段落E至F中任一项引述的方法，其中，确定在时间的与车辆相关联的估计位置至少包括：至少部分地基于车辆的附加系统和传感器数据的第一部分，确定与车辆相关联的参数；并且至少部分地基于附加误差模型和与车辆相关联的参数，确定在时间的与车辆相关联的估计位置，附加误差模型表示与附加系统相关联的误差概率。H: The method as recited in any of paragraphs E to F, wherein determining an estimated location associated with the vehicle at time comprises at least determining an estimate associated with the vehicle based at least in part on additional systems of the vehicle and the first portion of sensor data and determining an estimated position associated with the vehicle at time based at least in part on an additional error model and parameters associated with the vehicle, the additional error model representing an error probability associated with the additional system.

I：如段落E至H中任一项引述的方法，还包括：至少部分地基于参数，确定在时间的与物体相关联的附加估计位置；并且其中，确定在时间的与物体相关联的估计位置包括：至少部分地基于误差模型和与物体相关联的附加估计位置，确定在时间的与物体相关联的估计位置。I: The method as recited in any of paragraphs E to H, further comprising: determining, based at least in part on the parameters, additional estimated locations associated with the object at time; and wherein determining the estimates associated with the object at time The location includes determining an estimated location associated with the object at time based at least in part on the error model and additional estimated locations associated with the object.

J：如段落E至I中任一项引述的方法，其中，确定在时间的与物体相关联的估计位置包括：至少部分地基于误差模型和与物体相关联的参数，确定在时间的与物体相关联的估计位置的分布。J: The method as recited in any of paragraphs E to I, wherein determining an estimated location associated with the object at time comprises: determining an estimated location associated with the object at time based at least in part on an error model and parameters associated with the object Distribution of associated estimated locations.

K：如段落E至J中任一项引述的方法，其中，确定与车辆相关联的估计位置至少包括：至少部分地基于车辆的附加系统和传感器数据的第一部分，确定与车辆相关联的参数；并且至少部分地基于附加误差模型和与车辆相关联的参数，确定在该时间的与车辆相关联的估计位置的分布，附加误差模型表示与附加系统相关联的误差概率。K: The method as recited in any of paragraphs E to J, wherein determining an estimated location associated with the vehicle comprises at least determining parameters associated with the vehicle based at least in part on additional systems of the vehicle and the first portion of sensor data and determining a distribution of estimated positions associated with the vehicle at that time based at least in part on an additional error model representing an error probability associated with the additional system and parameters associated with the vehicle.

L：如段落E至K中任一项引述的方法，还包括：确定与车辆相关联的估计位置的分布和与物体相关联的估计位置的分布之间的重叠量；并且至少部分地基于重叠量来确定碰撞概率，并且其中，使车辆执行一个或多个动作是至少部分地基于碰撞概率的。L: The method as recited in any of paragraphs E to K, further comprising: determining an amount of overlap between the distribution of estimated locations associated with the vehicle and the distribution of estimated locations associated with the object; and based at least in part on the overlap amount to determine a probability of collision, and wherein causing the vehicle to perform one or more actions is based at least in part on the probability of collision.

M：如段落E至L中任一项引述的方法，还包括至少部分地基于参数来选择误差模型。M: The method as recited in any of paragraphs E to L, further comprising selecting an error model based at least in part on the parameters.

N：如段落E至M中任一项引述的方法，还包括：至少部分地基于车辆的附加系统和传感器数据的第二部分，确定与物体相关联的附加参数；并且至少部分地基于附加误差模型和与物体相关联的附加参数，确定与物体相关联的输出，附加误差模型表示与附加系统相关联的误差概率；并且其中，确定与物体相关联的参数包括：至少部分地基于车辆的系统和输出，确定与物体相关联的参数。N: The method as recited in any of paragraphs E to M, further comprising: determining additional parameters associated with the object based at least in part on additional systems of the vehicle and the second portion of the sensor data; and based at least in part on the additional errors a model and additional parameters associated with the object, determining an output associated with the object, the additional error model representing an error probability associated with the additional system; and wherein determining the parameter associated with the object comprises: a system based at least in part on the vehicle and output, which determine the parameters associated with the object.

O：如段落E至N中任一项引述的方法，其中，系统为感知系统并且附加系统为预测系统。O: The method as recited in any of paragraphs E to N, wherein the system is a perceptual system and the additional system is a predictive system.

P：如段落E至O中任一项引述的方法，还包括：至少部分地基于传感器数据的第一部分，确定在晚于时间的附加时间的与车辆相关联的附加估计位置；至少部分地基于车辆的系统和传感器数据的第二部分，确定与物体相关联的附加参数；至少部分地基于误差模型和与物体相关联的附加参数，确定在附加时间的与物体相关联的附加估计位置；并且至少部分地基于与车辆相关联的附加估计位置和与物体相关联的附加估计位置，使得车辆执行一个或多个动作。P: The method as recited in any of paragraphs E to O, further comprising: determining, based at least in part on the first portion of the sensor data, additional estimated locations associated with the vehicle at additional times later than time; based at least in part on a second portion of system and sensor data of the vehicle, determining additional parameters associated with the object; determining additional estimated locations associated with the object at additional times based at least in part on the error model and the additional parameters associated with the object; and The vehicle is caused to perform one or more actions based at least in part on the additional estimated locations associated with the vehicle and the additional estimated locations associated with the objects.

Q：如段落E至P中任一项引述的方法，还包括：至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置，确定碰撞概率；并且其中，使得车辆执行一个或多个动作包括：至少部分地基于碰撞概率，使得车辆进行改变速率或改变路线中的至少一个。Q: The method as recited in any of paragraphs E to P, further comprising: determining a probability of collision based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object; and wherein causing the vehicle to perform one or The plurality of actions include causing the vehicle to at least one of change rate or change route based at least in part on the probability of collision.

R：一种或多种非暂时性计算机可读介质，其存储指令，该指令当由一个或多个处理器执行时，使一个或多个计算设备执行操作，该操作包括：接收由与车辆相关联的传感器生成的传感器数据；至少部分地基于传感器数据的一部分，确定在该时间的与物体相关联的估计位置；至少部分地基于估计位置，从多个误差模型中确定误差模型；至少部分地基于误差模型和估计位置来确定与物体相关联的估计位置的分布；并且至少部分地基于估计位置的分布来确定用于导航车辆的一个或多个动作。R: One or more non-transitory computer-readable media that store instructions that, when executed by one or more processors, cause one or more computing devices to perform operations including: sensor data generated by an associated sensor; determining an estimated location associated with the object at the time based at least in part on a portion of the sensor data; determining an error model from a plurality of error models based at least in part on the estimated location; at least in part determining a distribution of estimated locations associated with the object based at least in part on the error model and the estimated locations; and determining one or more actions for navigating the vehicle based at least in part on the distribution of estimated locations.

S：如段落R中引述的一种或多种非暂时性计算机可读介质，操作还包括：至少部分地基于传感器数据的一部分来确定与车辆相关联的参数；至少部分地基于参数来确定估计位置，并且其中，误差模型与参数相关联。S: One or more non-transitory computer-readable media as recited in paragraph R, the operations further comprising: determining a parameter associated with the vehicle based at least in part on a portion of the sensor data; determining an estimate based at least in part on the parameter position, and where the error model is associated with the parameters.

T：如段落R或S中任一项引述的一种或多种非暂时性计算机可读介质，其中，确定误差模型进一步至少部分地基于以下各项中的一项或多项：物体的分类、物体的速度、物体的大小、环境中物体的数量、环境中的天气状况、一天中的某个时间、或一年中的某个时间。T: One or more non-transitory computer-readable media as recited in any of paragraphs R or S, wherein determining the error model is further based, at least in part, on one or more of the following: classification of objects , the speed of the object, the size of the object, the number of objects in the environment, the weather conditions in the environment, the time of day, or the time of the year.

U：一种自动驾驶车辆，包括：一个或多个传感器；一个或多个处理器；以及一种或多种计算机可读介质，其存储指令，该指令当由一个或多个处理器执行时，使一个或多个处理器执行操作，操作包括：获取由一个或多个传感器生成的传感器数据；至少部分地基于传感器数据的第一部分确定自动驾驶车辆的估计位置；至少部分地基于传感器数据的第二部分确定物体的估计位置；确定与物体的估计位置相关联的不确定性模型；至少部分地基于不确定性模型和物体的估计位置来确定与物体相关联的估计位置的分布；至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置的概率来确定自动驾驶车辆和物体之间的碰撞概率；并且至少部分地基于碰撞概率使自动驾驶车辆执行一个或多个动作。U: An autonomous vehicle comprising: one or more sensors; one or more processors; and one or more computer-readable media storing instructions, which when executed by the one or more processors , causing one or more processors to perform operations comprising: obtaining sensor data generated by one or more sensors; determining an estimated location of the autonomous vehicle based at least in part on a first portion of the sensor data; A second part determines an estimated location of the object; determines an uncertainty model associated with the estimated location of the object; determines a distribution of estimated locations associated with the object based at least in part on the uncertainty model and the estimated location of the object; at least in part determining a probability of collision between the autonomous vehicle and the object based at least in part on the estimated position associated with the vehicle and the estimated position associated with the object; and causing the autonomous vehicle to perform one or more actions based at least in part on the probability of collision.

V：如段落U中引述的自动驾驶车辆，操作还包括：确定与附加系统确定自动驾驶车辆的估计位置相关联的附加不确定性模型；并且至少部分地基于附加不确定性模型和自动驾驶车辆的估计位置，确定与自动驾驶车辆相关联的估计位置的概率，并且其中，确定自动驾驶车辆与物体之间的碰撞概率至少包括：确定与自动驾驶车辆相关联的估计位置的概率和与物体相关联的估计位置的概率之间的重叠量；并且至少部分地基于重叠量来确定碰撞概率。V: The autonomous vehicle as recited in paragraph U, the operations further comprising: determining an additional uncertainty model associated with the additional system determining the estimated location of the autonomous vehicle; and based at least in part on the additional uncertainty model and the autonomous vehicle , determining the probability of the estimated position associated with the autonomous vehicle, and wherein determining the probability of collision between the autonomous vehicle and the object includes at least: determining the probability of the estimated position associated with the autonomous vehicle and the correlation with the object and determining a probability of collision based at least in part on the amount of overlap.

W：如段落U或V中引述的自动驾驶车辆，其中：至少部分地基于自动驾驶车辆的附加系统进一步确定物体的估计位置；操作还包括确定与附加系统确定物体的估计位置相关联的附加不确定性模型；并且至少部分地基于附加不确定性模型进一步确定估计位置的概率。W: An autonomous vehicle as recited in paragraphs U or V, wherein: the estimated position of the object is further determined based at least in part on an additional system of the autonomous vehicle; the operations further include determining additional parameters associated with the additional system determining the estimated position of the object a deterministic model; and further determining a probability of the estimated location based at least in part on the additional uncertainty model.

X：一种方法，包括：从车辆的一个或多个传感器接收传感器数据；至少部分地基于传感器数据的第一部分来确定与车辆相关联的估计位置；至少部分地基于车辆的系统和传感器数据的第二部分来确定与物体相关联的参数；确定与系统确定与物体相关联的参数相关联的不确定性模型；至少部分地基于与物体相关联的参数和不确定性模型来确定与物体相关联的估计位置；并且至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置，使车辆执行一个或多个动作。X: A method comprising: receiving sensor data from one or more sensors of a vehicle; determining an estimated location associated with the vehicle based at least in part on a first portion of the sensor data; based at least in part on systems and sensor data of the vehicle a second part to determine a parameter associated with the object; determine an uncertainty model associated with the system determining the parameter associated with the object; determine an uncertainty model associated with the object based at least in part on the parameter associated with the object and the uncertainty model and causing the vehicle to perform one or more actions based at least in part on the estimated position associated with the vehicle and the estimated position associated with the object.

Y：如段落X中引述的方法，还包括从一个或多个计算设备接收不确定性模型，不确定性模型是至少部分地基于由一个或多个车辆生成的传感器数据生成的。Y: The method as recited in paragraph X, further comprising receiving an uncertainty model from one or more computing devices, the uncertainty model being generated based at least in part on sensor data generated by the one or more vehicles.

Z：如段落X或Y中任一项引述的方法，其中，确定与物体相关联的参数包括至少部分地基于系统和传感器数据的第二部分来确定以下各项中的至少一项：与物体相关联的物体类型；物体在环境内的位置；物体的速度；或物体在环境内的行进方向。Z: The method as recited in any of paragraphs X or Y, wherein determining a parameter associated with the object comprises determining at least one of: an association with the object based at least in part on the second portion of the system and sensor data The associated object type; the object's position within the environment; the object's velocity; or the object's direction of travel within the environment.

AA：如段落X至Z中任一项引述的方法，其中，确定与车辆相关联的估计位置至少包括：至少部分地基于车辆的附加系统和传感器数据的第一部分来确定与车辆相关联的参数；确定与附加系统确定与车辆相关联的参数相关联的附加不确定性模型；并且至少部分地基于与车辆相关联的参数和附加不确定性模型来确定与车辆相关联的估计位置。AA: The method as recited in any of paragraphs X to Z, wherein determining an estimated position associated with the vehicle comprises at least determining parameters associated with the vehicle based at least in part on additional systems of the vehicle and the first portion of sensor data determining an additional uncertainty model associated with the additional system determining parameters associated with the vehicle; and determining an estimated location associated with the vehicle based at least in part on the parameters associated with the vehicle and the additional uncertainty model.

AB：如段落X至AA中任一项引述的方法，还包括：至少部分地基于参数确定与物体相关联的附加估计位置，并且其中，确定与物体相关联的估计位置包括至少部分地基于与物体相关联的附加估计位置和不确定性模型来确定与物体相关联的估计位置。AB: The method as recited in any of paragraphs X to AA, further comprising: determining additional estimated positions associated with the object based at least in part on the parameters, and wherein determining the estimated positions associated with the object comprises Additional estimated positions associated with the objects and uncertainty models to determine estimated positions associated with the objects.

AC：如段落X至AB中任一项引述的方法，其中，确定与物体相关联的估计位置包括至少部分地基于与物体相关联的参数和不确定性模型来确定与物体相关联的估计位置的分布。AC: The method as recited in any of paragraphs X to AB, wherein determining an estimated location associated with the object comprises determining an estimated location associated with the object based at least in part on parameters associated with the object and an uncertainty model Distribution.

AD：如段落X至AC中任一项引述的方法，其中，确定与车辆相关联的估计位置至少包括：至少部分地基于车辆的附加系统和传感器数据的第一部分来确定与车辆相关联的参数；确定与附加系统确定与车辆相关联的参数相关联的附加不确定性模型；并且至少部分地基于与车辆相关联的参数和附加不确定性模型来确定与车辆相关联的估计位置的分布。AD: The method as recited in any of paragraphs X to AC, wherein determining an estimated position associated with the vehicle comprises at least determining parameters associated with the vehicle based at least in part on additional systems of the vehicle and the first portion of sensor data determining an additional uncertainty model associated with the additional system determining parameters associated with the vehicle; and determining a distribution of estimated locations associated with the vehicle based at least in part on the parameters associated with the vehicle and the additional uncertainty model.

AE：如段落X至AD中任一项引述的方法，还包括：确定与车辆相关联的估计位置的分布和与物体相关联的估计位置的分布之间的重叠量；并且至少部分地基于重叠量来确定碰撞概率，并且其中，至少部分地基于碰撞概率使车辆执行一个或多个动作。AE: The method as recited in any of paragraphs X to AD, further comprising: determining an amount of overlap between the distribution of estimated locations associated with the vehicle and the distribution of estimated locations associated with the object; and based at least in part on the overlap amount to determine a probability of collision, and wherein the vehicle is caused to perform one or more actions based at least in part on the probability of collision.

AF：如段落X至AE中任一项引述的方法，还包括：至少部分地基于车辆的附加系统和传感器数据的第三部分，确定与物体相关联的附加参数；并且确定与附加系统确定与物体相关联的附加参数相关联的附加不确定性模型；并且其中，确定与物体相关联的估计位置进一步至少部分地基于附加参数和附加不确定性模型。AF: The method as recited in any of paragraphs X to AE, further comprising: determining an additional parameter associated with the object based at least in part on an additional system of the vehicle and the third portion of the sensor data; and determining with the additional system determining an additional parameter associated with the object; an additional uncertainty model associated with additional parameters associated with the object; and wherein determining an estimated location associated with the object is further based at least in part on the additional parameters and the additional uncertainty model.

AG：如段落X至AF中任一项引述的方法，还包括：至少部分地基于车辆的附加系统和传感器数据的第二部分来确定与物体相关联的附加参数；确定与附加系统确定与物体相关联的附加参数相关联的附加不确定性模型；并且至少部分地基于与物体相关联的附加参数和附加不确定性模型来确定与物体相关联的输出，并且其中，确定与物体相关联的参数包括至少部分地基于车辆的系统和输出来确定与物体相关联的参数。AG: The method as recited in any of paragraphs X to AF, further comprising: determining an additional parameter associated with the object based at least in part on an additional system of the vehicle and the second portion of the sensor data; determining that the additional system determines an additional parameter associated with the object; an additional uncertainty model associated with an associated additional parameter; and determining an output associated with the object based at least in part on the additional parameter associated with the object and the additional uncertainty model, and wherein determining an output associated with the object The parameters include determining parameters associated with the object based at least in part on systems and outputs of the vehicle.

AH：如段落X至AG中任一项引述的方法，还包括：至少部分地基于车辆的系统和传感器数据的第三部分来确定与附加物体相关联的参数；确定与系统确定与附加物体相关联的参数相关联的附加不确定性模型；至少部分地基于与附加物体相关联的参数和附加不确定性模型来确定与附加物体相关联的估计位置；并且其中，使车辆执行一个或多个动作进一步至少部分地基于与附加物体相关联的估计位置。AH: The method as recited in any of paragraphs X to AG, further comprising: determining a parameter associated with the additional object based at least in part on the system of the vehicle and the third portion of the sensor data; determining a parameter associated with the additional object determined by the system; an additional uncertainty model associated with the associated parameter; determining an estimated location associated with the additional object based at least in part on the parameter associated with the additional object and the additional uncertainty model; and wherein the vehicle is caused to perform one or more of the The action is further based, at least in part, on the estimated location associated with the additional object.

AI：如段落X至AH中任一项引述的方法，还包括：至少部分地基于与车辆相关联的估计位置和与物体相关联的估计位置来确定碰撞概率，并且其中，使车辆执行一个或多个动作包括至少部分地基于碰撞概率使车辆进行改变速率或改变路线中的至少一个。AI: The method as recited in any of paragraphs X to AH, further comprising: determining a probability of collision based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object, and wherein causing the vehicle to perform one of or The plurality of actions include causing the vehicle to at least one of change rate or change route based at least in part on the probability of collision.

AJ：一种或多种非暂时性计算机可读介质，其存储指令，该指令当由一个或多个处理器执行时，使一个或多个计算设备执行操作，操作包括：接收由与车辆相关联的传感器生成的传感器数据；至少部分地基于传感器数据的一部分，确定与物体相关联的估计位置；至少部分地基于估计位置，从多个不确定性模型中确定不确定性模型；至少部分地基于不确定性模型和估计位置来确定与物体相关联的估计位置的分布；并且至少部分地基于估计位置的分布来确定用于导航车辆的一个或多个动作。AJ: One or more non-transitory computer-readable media that store instructions that, when executed by one or more processors, cause one or more computing devices to perform operations including: sensor data generated by an associated sensor; determining an estimated location associated with the object based at least in part on a portion of the sensor data; determining an uncertainty model from a plurality of uncertainty models based at least in part on the estimated location; A distribution of estimated locations associated with the object is determined based on the uncertainty model and the estimated locations; and one or more actions for navigating the vehicle are determined based at least in part on the distribution of estimated locations.

AK：如段落AJ引述的一种或多种非暂时性计算机可读介质，操作还包括：至少部分地基于传感器数据的一部分来确定与车辆相关联的参数；至少部分地基于参数来确定估计位置，并且其中，不确定性模型与参数相关联。AK: One or more non-transitory computer-readable media as recited in paragraph AJ, the operations further comprising: determining a parameter associated with the vehicle based at least in part on a portion of the sensor data; determining an estimated location based at least in part on the parameter , and where the uncertainty model is associated with the parameter.

AL：如段落AJ或AK中任一项引述的一种或多种非暂时性计算机可读介质，操作还包括：至少部分地基于传感器数据的附加部分来确定与车辆相关联的估计位置；至少部分地基于估计位置，从多个不确定性模型中确定附加不确定性模型；并且至少部分地基于附加不确定性模型和与车辆相关联的估计位置，确定与车辆相关联的估计位置的分布，并且其中，确定一个或多个动作进一步至少部分地基于与车辆相关联的估计位置的分布。AL: One or more non-transitory computer-readable media as recited in any of paragraphs AJ or AK, the operations further comprising: determining an estimated location associated with the vehicle based at least in part on the additional portion of the sensor data; at least determining an additional uncertainty model from the plurality of uncertainty models based in part on the estimated location; and determining a distribution of estimated locations associated with the vehicle based at least in part on the additional uncertainty model and the estimated location associated with the vehicle , and wherein determining the one or more actions is further based, at least in part, on a distribution of estimated locations associated with the vehicle.

AM：如段落AJ至AL中任一项引述的一种或多种非暂时性计算机可读介质，操作还包括：至少部分地基于与车辆相关联的估计位置的分布和与物体相关联的估计位置的分布来确定碰撞概率，并且其中，至少部分地基于碰撞概率来确定一个或多个动作。AM: One or more non-transitory computer-readable media as recited in any of paragraphs AJ to AL, the operations further comprising: based at least in part on a distribution of estimated locations associated with vehicles and estimates associated with objects A distribution of locations to determine a probability of collision, and wherein the one or more actions are determined based at least in part on the probability of collision.

AN：如段落AJ至AM中任一项引述的一种或多种非暂时性计算机可读介质，其中，确定不确定性模型进一步至少部分地基于以下各项中的一项或多项：物体的分类、物体的速度、物体的大小、环境中物体的数量、环境中的天气状况、一天中的某个时间、或一年中的某个时间。AN: One or more non-transitory computer-readable media as recited in any of paragraphs AJ to AM, wherein determining the uncertainty model is further based, at least in part, on one or more of the following: an object classification of objects, the speed of objects, the size of objects, the number of objects in the environment, the weather conditions in the environment, the time of day, or the time of year.

Claims (15)

Translated fromChinese

1.一种自动驾驶车辆，包括：1. An autonomous vehicle comprising:一个或多个传感器；one or more sensors;一个或多个处理器；以及one or more processors; and一种或多种计算机可读介质，其存储指令，所述指令当由所述一个或多个处理器执行时，使得所述一个或多个处理器执行操作，所述操作包括：One or more computer-readable media storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:从所述一个或多个传感器获取传感器数据；obtaining sensor data from the one or more sensors;至少部分地基于所述传感器数据的第一部分，确定所述自动驾驶车辆在未来时间的估计位置；determining an estimated location of the autonomous vehicle at a future time based at least in part on the first portion of the sensor data;至少部分地基于所述自动驾驶车辆的系统和所述传感器数据的第二部分，确定物体在所述未来时间的估计位置；determining an estimated location of the object at the future time based at least in part on the system of the autonomous vehicle and the second portion of the sensor data;至少部分地基于误差模型和所述物体的估计位置，确定与所述物体相关联的估计位置的分布，所述误差模型表示与所述系统相关联的误差概率；determining a distribution of estimated positions associated with the object based at least in part on an error model representing an error probability associated with the system and the estimated position of the object;至少部分地基于所述自动驾驶车辆的估计位置以及与所述物体相关联的所述估计位置的分布，确定所述自动驾驶车辆和所述物体之间的碰撞概率；并且determining a probability of collision between the autonomous vehicle and the object based at least in part on the estimated location of the autonomous vehicle and the distribution of the estimated locations associated with the object; and至少部分地基于所述碰撞概率，使得所述自动驾驶车辆执行一个或多个动作。The autonomous vehicle is caused to perform one or more actions based at least in part on the probability of collision.2.根据权利要求1所述的自动驾驶车辆，所述操作还包括从一个或多个计算设备接收所述误差模型，所述误差模型是至少使用由一个或多个车辆生成的传感器数据来生成的。2. The autonomous vehicle of claim 1, the operations further comprising receiving the error model from one or more computing devices, the error model generated using at least sensor data generated by one or more vehicles of.3.根据权利要求1或2所述的自动驾驶车辆，所述操作还包括：3. The autonomous vehicle of claim 1 or 2, the operations further comprising:至少部分地基于附加误差模型和所述车辆的估计位置，确定与所述自动驾驶车辆相关联的估计位置的分布；determining a distribution of estimated positions associated with the autonomous vehicle based at least in part on the additional error model and the estimated positions of the vehicles;并且其中，确定所述自动驾驶车辆和所述物体之间的所述碰撞概率至少包括：And wherein, determining the collision probability between the autonomous vehicle and the object at least includes:确定与所述自动驾驶车辆相关联的所述估计位置的分布和与所述物体相关联的所述估计位置的分布之间的重叠量；并且determining an amount of overlap between the distribution of the estimated locations associated with the autonomous vehicle and the distribution of the estimated locations associated with the object; and至少部分地基于所述重叠量，确定所述碰撞概率。The collision probability is determined based at least in part on the amount of overlap.4.根据权利要求1-3中任一项所述的自动驾驶车辆，其中：4. The autonomous vehicle of any one of claims 1-3, wherein:至少部分地基于所述自动驾驶车辆的附加系统，进一步确定所述物体在所述未来时间的所述估计位置；并且further determining the estimated location of the object at the future time based at least in part on an additional system of the autonomous vehicle; and至少部分地基于附加误差模型，进一步确定所述估计位置的分布，所述附加误差模型表示与所述附加系统相关联的误差分布。The distribution of the estimated positions is further determined based at least in part on an additional error model, the additional error model representing an error distribution associated with the additional system.5.一种方法，包括：5. A method comprising:从车辆的一个或多个传感器接收传感器数据；receive sensor data from one or more sensors of the vehicle;至少部分地基于所述传感器数据的第一部分，确定在一时间的与所述车辆相关联的估计位置；determining an estimated location associated with the vehicle at a time based at least in part on the first portion of the sensor data;至少部分地基于所述车辆的系统和所述传感器数据的第二部分，确定与物体相关联的参数；determining a parameter associated with the object based at least in part on the system of the vehicle and the second portion of the sensor data;至少部分地基于误差模型和与所述物体相关联的所述参数，确定在所述时间的与所述物体相关联的估计位置，所述误差模型表示与所述系统相关联的误差概率；并且determining an estimated location associated with the object at the time based at least in part on an error model representing an error probability associated with the system and the parameter associated with the object; and至少部分地基于与所述车辆相关联的所述估计位置和与所述物体相关联的所述估计位置，使得所述车辆执行一个或多个动作。The vehicle is caused to perform one or more actions based at least in part on the estimated position associated with the vehicle and the estimated position associated with the object.6.根据权利要求5项所述的方法，还包括从一个或多个计算设备接收所述误差模型，所述误差模型是至少使用由一个或多个车辆生成的传感器数据来生成的。6. The method of claim 5, further comprising receiving the error model from one or more computing devices, the error model generated using at least sensor data generated by one or more vehicles.7.根据权利要求5或6所述的方法，其中，所述参数包括以下各项中的至少一项：7. The method of claim 5 or 6, wherein the parameter comprises at least one of the following:与所述物体相关联的物体类型；the object type associated with the object;所述物体在环境内的位置；the location of the object within the environment;所述物体的速度；或the speed of the object; or所述物体在所述环境内的行进方向。The direction of travel of the object within the environment.8.根据权利要求5-7中任一项所述的方法，其中，确定在所述时间的与所述车辆相关联的所述估计位置至少包括：8. The method of any of claims 5-7, wherein determining the estimated location associated with the vehicle at the time comprises at least:至少部分地基于所述车辆的附加系统和所述传感器数据的第一部分，确定与所述车辆相关联的参数；并且determining parameters associated with the vehicle based at least in part on additional systems of the vehicle and the first portion of the sensor data; and至少部分地基于附加误差模型和与所述车辆相关联的所述参数，确定在所述时间的与所述车辆相关联的所述估计位置，所述附加误差模型表示与所述附加系统相关联的误差概率。determining the estimated position associated with the vehicle at the time based, at least in part, on an additional error model representing an association with the additional system and the parameter associated with the vehicle error probability.9.根据权利要求5-8中任一项所述的方法，还包括：9. The method of any one of claims 5-8, further comprising:至少部分地基于所述参数，确定在所述时间的与所述物体相关联的附加估计位置；determining additional estimated locations associated with the object at the time based at least in part on the parameters;并且其中，确定在所述时间的与所述物体相关联的所述估计位置包括：至少部分地基于所述误差模型和与所述物体相关联的所述附加估计位置，确定在所述时间的与所述物体相关联的所述估计位置。and wherein determining the estimated location associated with the object at the time includes determining an estimated location associated with the object at the time based at least in part on the error model and the additional estimated location associated with the object the estimated location associated with the object.10.根据权利要求5-9中任一项所述的方法，其中，确定在所述时间的与所述物体相关联的所述估计位置包括：至少部分地基于所述误差模型和与所述物体相关联的所述参数，确定在所述时间的与所述物体相关联的估计位置的分布。10. The method of any of claims 5-9, wherein determining the estimated location associated with the object at the time comprises: based at least in part on the error model and a correlation with the The parameter associated with the object determines a distribution of estimated positions associated with the object at the time.11.根据权利要求5-10中任一项所述的方法，还包括至少部分地基于所述参数来选择所述误差模型。11. The method of any of claims 5-10, further comprising selecting the error model based at least in part on the parameters.12.根据权利要求5-11中任一项所述的方法，还包括：12. The method of any one of claims 5-11, further comprising:至少部分地基于所述车辆的附加系统和所述传感器数据的第二部分，确定与所述物体相关联的附加参数；并且determining additional parameters associated with the object based at least in part on additional systems of the vehicle and the second portion of the sensor data; and至少部分地基于附加误差模型和与所述物体相关联的所述附加参数，确定与所述物体相关联的输出，所述附加误差模型表示与所述附加系统相关联的误差概率；determining an output associated with the object based at least in part on an additional error model representing an error probability associated with the additional system and the additional parameter associated with the object;并且其中，确定与所述物体相关联的所述参数包括：至少部分地基于所述车辆的系统和所述输出，确定与所述物体相关联的所述参数。And wherein determining the parameter associated with the object includes determining the parameter associated with the object based at least in part on a system of the vehicle and the output.13.根据权利要求5-11中任一项所述的方法，还包括：13. The method of any one of claims 5-11, further comprising:至少部分地基于所述传感器数据的第一部分，确定在晚于所述时间的附加时间的与所述车辆相关联的附加估计位置；determining additional estimated locations associated with the vehicle at additional times later than the time based at least in part on the first portion of the sensor data;至少部分地基于所述车辆的系统和所述传感器数据的第二部分，确定与所述物体相关联的附加参数；determining additional parameters associated with the object based at least in part on the systems of the vehicle and the second portion of the sensor data;至少部分地基于所述误差模型和与所述物体相关联的所述附加参数，确定在所述附加时间的与所述物体相关联的附加估计位置；并且determining additional estimated locations associated with the object at the additional times based at least in part on the error model and the additional parameters associated with the object; and至少部分地基于与所述车辆相关联的所述附加估计位置和与所述物体相关联的所述附加估计位置，使得所述车辆执行一个或多个动作。The vehicle is caused to perform one or more actions based at least in part on the additional estimated location associated with the vehicle and the additional estimated location associated with the object.14.根据权利要求5-13中任一项所述的方法，还包括：14. The method of any one of claims 5-13, further comprising:至少部分地基于与所述车辆相关联的所述估计位置和与所述物体相关联的所述估计位置，确定碰撞概率；determining a probability of collision based at least in part on the estimated location associated with the vehicle and the estimated location associated with the object;并且其中，使得所述车辆执行所述一个或多个动作包括：至少部分地基于所述碰撞概率，使得所述车辆进行改变速率或改变路线中的至少一个。And wherein causing the vehicle to perform the one or more actions includes causing the vehicle to perform at least one of a rate change or a route change based at least in part on the collision probability.15.一种或多种非暂时性计算机可读介质，其存储指令，所述指令当被执行时，使得一个或多个处理器执行根据权利要求5-14中任一项所述的方法。15. One or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform the method of any of claims 5-14.

Images