CN113246968A - Automatic driving automobile accurate parking control method and device - Google Patents

Abstract

Translated fromChinese

本发明公开了一种自动驾驶汽车精准停车控制方法及装置，其方法包括：在自动驾驶汽车停车期间，所述自动驾驶汽车分别获取当前时间的距离值、角速度值以及加速度值；所述自动驾驶汽车将所述距离值、角速度值以及加速度值输入到已训练好的冈波茨模型中，得到当前时间的轨迹曲线；所述自动驾驶汽车利用所述当前时间的轨迹曲线，分别计算待移动的曲线距离值和角度，以便根据所述曲线距离值和角度进行精准停车。

The invention discloses a precise parking control method and device for an automatic driving vehicle. The method includes: during the parking period of the automatic driving vehicle, the automatic driving vehicle obtains the distance value, angular velocity value and acceleration value of the current time respectively; the automatic driving The car inputs the distance value, angular velocity value and acceleration value into the trained Gangpotz model to obtain the trajectory curve of the current time; the self-driving car uses the trajectory curve of the current time to calculate the to-be-moved trajectory curve respectively. Curve distance value and angle for precise parking according to said curve distance value and angle.

Description

Translated fromChinese

一种自动驾驶汽车精准停车控制方法及装置A method and device for precise parking control of an autonomous vehicle

技术领域technical field

本发明涉及计算机软件技术领域，特别涉及一种自动驾驶汽车精准停车控制方法及装置。The invention relates to the technical field of computer software, in particular to a method and device for precise parking control of an automatic driving vehicle.

背景技术Background technique

随着人们生活质量的日益提高，汽车出行成为人们生活中不可或缺的一个重要组成部分，汽车带给人们的不仅仅是便捷，同时还有时间的节省。但是在现实生活中，人们会由于工作的繁忙、熬夜或者长时间驾驶等情况而感到疲劳，这时驾驶汽车将会很容易发生交通事故。更甚至有一些不重视交通规则的人酒醉驾驶，在驾驶时接打电话以及走神等更是对他人以及自身都造成了严重的生命威胁。而正因为这些不遵守道路交通规则的驾驶员给我们的道路和交通环境带来了巨大的压力，人们的出行尤其是高峰出行的效率被极大的降低。为了减少交通事故率和道路的负担。在1970年左右，美国等发达国家开始将人工智能运用于汽车上，因而研发出自动驾驶汽车，而随着大数据的发展，自动驾驶汽车更是被推向了一个新的高度。With the increasing improvement of people's quality of life, car travel has become an indispensable and important part of people's lives. Cars bring people not only convenience, but also time savings. However, in real life, people will feel fatigued due to busy work, staying up late or driving for a long time. At this time, driving a car will be prone to traffic accidents. There are even some people who don't pay attention to traffic rules and drive while drunk, making phone calls and distracted while driving, which has caused serious life-threatening to others and themselves. And it is precisely because these drivers who do not obey the road traffic rules have brought tremendous pressure to our roads and traffic environment, and the efficiency of people's travel, especially peak travel, has been greatly reduced. In order to reduce the traffic accident rate and the burden on the road. Around 1970, developed countries such as the United States began to apply artificial intelligence to cars, thus developing self-driving cars. With the development of big data, self-driving cars have been pushed to a new height.

而在自动驾驶汽车的研究中，对自动驾驶汽车的高效平行停车的研究一直都是一个难点，尤其是实现快速且避障的自动停车，更是当前研究学者们的研究重点和难题。目前已经有的自动驾驶汽车停车控制系统有例如：Paromtchik提出的基于超声波测距数据处理和正弦函数控制车辆自动停车的方法、M.Wada提出的基于路径规划和人机界面的多级驾驶员辅助停车控制系统、利用模糊逻辑实现了车辆停车控制等。这些方法在数据存储方面，不需要耗费大量的空间对数据进行存储处理，因此存储开销。但是需要大量的非线性方程的计算，计算时间复杂度极高，导致控制汽车反应很慢，其次很不容易在实践中进行实现，很难让自动驾驶汽车实现稳定的自动避障停车。In the research of autonomous vehicles, the research on efficient parallel parking of autonomous vehicles has always been a difficult point, especially the realization of fast and obstacle-avoiding automatic parking is the focus and problem of current researchers. At present, there are automatic parking control systems for autonomous vehicles, such as: Paromtchik's method based on ultrasonic ranging data processing and sine function to control the vehicle's automatic parking, M.Wada's multi-level driver assistance based on path planning and human-machine interface Parking control system, using fuzzy logic to realize vehicle parking control, etc. In terms of data storage, these methods do not need to consume a large amount of space for data storage processing, so storage overhead is incurred. However, a large number of nonlinear equations are required for calculation, and the calculation time complexity is extremely high, resulting in slow response of the control car. Secondly, it is not easy to implement in practice, and it is difficult for autonomous vehicles to achieve stable automatic obstacle avoidance parking.

现有的自动驾驶汽车自动停车控制系统，主要有利用模糊逻辑来实现控制的方法，该方法利用模糊逻辑模型，结合多种传感器所接受到的信息，将多个传感器的数据进行融合并放入模型中进行运算，从而来控制自动驾驶汽车的停车。The existing automatic parking control systems for autonomous vehicles mainly use fuzzy logic to achieve control. Operations are performed in the model to control the parking of the self-driving car.

现在自动驾驶汽车所使用的模糊逻辑控制方法，在复杂系统的语言描述场景下十分适用，同时它可以用来制定和翻译语言表达的人类经验，以适当的自动控制策略。但是也有其缺陷，就是该方法在使用时，的算法时间复杂度高，而且算法的稳定性强，因此在实际运用测试中，我们实验的自动驾驶汽车会出现反应慢、卡顿或者与停车位其他汽车有碰撞的情况。The fuzzy logic control method currently used in self-driving cars is very suitable in the context of language description of complex systems, and it can be used to formulate and translate human experience expressed in language to appropriate automatic control strategies. However, it also has its drawbacks, that is, when this method is used, the algorithm has a high time complexity, and the algorithm has strong stability. Therefore, in the actual application test, the self-driving car we experimented will appear slow in response, stuck or inconsistent with the parking space. Other cars have had collisions.

发明内容SUMMARY OF THE INVENTION

根据本发明实施例提供的方案解决的技术问题是如何在保证自动驾驶汽车停车时间最短的情况下，规划出最优的避障停车轨迹，并且对生成的候选轨迹进行了碰撞检测。The technical problem solved by the solution provided by the embodiment of the present invention is how to plan an optimal obstacle avoidance parking trajectory while ensuring the shortest parking time of the autonomous driving vehicle, and perform collision detection on the generated candidate trajectory.

根据本发明实施例提供的一种自动驾驶汽车精准停车控制方法，包括：A method for precise parking control of an autonomous vehicle provided according to an embodiment of the present invention includes:

在自动驾驶汽车停车期间，所述自动驾驶汽车分别获取当前时间的距离值、角速度值以及加速度值；During the parking period of the self-driving car, the self-driving car obtains the distance value, angular velocity value and acceleration value of the current time respectively;

所述自动驾驶汽车将所述距离值、角速度值以及加速度值输入到已训练好的冈波茨模型中，得到当前时间的轨迹曲线；The self-driving car inputs the distance value, the angular velocity value and the acceleration value into the trained Gangpotz model to obtain the trajectory curve of the current time;

所述自动驾驶汽车利用所述当前时间的轨迹曲线，分别计算待移动的曲线距离值和角度，以便根据所述曲线距离值和角度进行精准停车。The self-driving car uses the trajectory curve of the current time to separately calculate the distance value and angle of the curve to be moved, so as to accurately stop according to the curve distance value and angle.

根据本发明实施例提供的一种自动驾驶汽车精准停车控制装置，包括：A precise parking control device for an autonomous vehicle provided according to an embodiment of the present invention includes:

获取模块，用于在自动驾驶汽车停车期间，分别获取当前时间的距离值、角速度值以及加速度值；The acquisition module is used to acquire the distance value, angular velocity value and acceleration value of the current time respectively during the parking period of the autonomous vehicle;

输入模块，用于将所述距离值、角速度值以及加速度值输入到已训练好的冈波茨模型中，得到当前时间的轨迹曲线；an input module for inputting the distance value, the angular velocity value and the acceleration value into the trained Gangpotz model to obtain the trajectory curve of the current time;

计算及停车模块，用于利用所述当前时间的轨迹曲线，分别计算待移动的曲线距离值和角度，以便根据所述曲线距离值和角度进行精准停车。The calculation and parking module is used to calculate the distance value and angle of the curve to be moved by using the trajectory curve of the current time, so as to perform accurate parking according to the curve distance value and angle.

根据本发明实施例提供的方案，使得自动驾驶汽车能够快速的进行平行位的停车，同时能避免和周围的障碍物或者其它汽车进行碰撞或剐蹭。According to the solution provided by the embodiment of the present invention, the autonomous vehicle can quickly park in parallel positions, and at the same time, it can avoid collision or scratch with surrounding obstacles or other vehicles.

附图说明Description of drawings

此处所说明的附图用来提供对本发明的进一步理解，构成本发明的一部分，本发明的示意性实施例及其说明用于理解本发明，并不构成对本发明的不当限定。在附图中：The accompanying drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to understand the present invention and do not constitute an improper limitation of the present invention. In the attached image:

图1是本发明实施例提供的一种自动驾驶汽车精准停车控制方法的流程图；1 is a flowchart of a method for precise parking control of an autonomous vehicle provided by an embodiment of the present invention;

图2是本发明实施例提供的一种自动驾驶汽车精准停车控制装置的示意图。FIG. 2 is a schematic diagram of a precise parking control device for an automatic driving vehicle according to an embodiment of the present invention.

具体实施方式Detailed ways

以下结合附图对本发明的优选实施例进行详细说明，应当理解，以下所说明的优选实施例仅用于说明和解释本发明，并不用于限定本发明。The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the preferred embodiments described below are only used to illustrate and explain the present invention, but not to limit the present invention.

图1是本发明实施例提供的一种自动驾驶汽车精准停车控制方法的流程图，如图1所示，包括：1 is a flowchart of a method for precise parking control of an autonomous vehicle provided by an embodiment of the present invention, as shown in FIG. 1 , including:

步骤S101：在自动驾驶汽车停车期间，所述自动驾驶汽车分别获取当前时间的距离值、角速度值以及加速度值；Step S101: During the parking period of the self-driving car, the self-driving car obtains the distance value, angular velocity value and acceleration value of the current time respectively;

步骤S102：所述自动驾驶汽车将所述距离值、角速度值以及加速度值输入到已训练好的冈波茨模型中，得到当前时间的轨迹曲线；Step S102: the self-driving car inputs the distance value, angular velocity value and acceleration value into the trained Gangpotz model to obtain the trajectory curve of the current time;

步骤S103：所述自动驾驶汽车利用所述当前时间的轨迹曲线，分别计算待移动的曲线距离值和角度，以便根据所述曲线距离值和角度进行精准停车。Step S103: The self-driving car uses the trajectory curve of the current time to calculate the distance value and angle of the curve to be moved respectively, so as to accurately stop according to the curve distance value and angle.

其中，所述距离值包括以下任一或组合：前左车轮距离前障碍物或前停车线的第一距离值；前右车轮距离前障碍物或前停车线的第二距离值；后左车轮距离后障碍物或后停车线的第三距离值；后右车轮距离后障碍物或后停车线的第四距离值。所述自动驾驶汽车分别获取当前时间的距离值、角速度值以及加速度值包括：通过所述自动驾驶汽车的摄像头和激光距离扫描仪测量当前时间的距离值；通过所述自动驾驶汽车的单轴陀螺仪传感器获取当前时间的角速度值；通过所述自动驾驶汽车的三维加速度传感器测量当前时间的加速度值。Wherein, the distance value includes any one or a combination of the following: the first distance value of the front left wheel from the front obstacle or the front stop line; the second distance value of the front right wheel from the front obstacle or the front stop line; the rear left wheel The third distance value from the rear obstacle or the rear stop line; the fourth distance value of the rear right wheel from the rear obstacle or the rear stop line. The self-driving car obtains the distance value, angular velocity value and acceleration value of the current time respectively including: measuring the distance value of the current time through the camera and laser distance scanner of the self-driving car; using the single-axis gyro of the self-driving car The angular velocity value of the current time is obtained by the instrument sensor; the acceleration value of the current time is measured by the three-dimensional acceleration sensor of the self-driving car.

具体地说，所述已训练好的冈波茨模型包括：所述自动驾驶汽车分别获取所述自动驾驶汽车的行驶速度、后车轮中点的行驶距离以及可用停车空间的总长度值；所述自动驾驶汽车利用所述自动驾驶汽车的行驶速度、后车轮中点的行驶距离以及可用停车空间的总长度值，对初始冈波茨模型进行训练，得到已训练好的冈波茨模型。Specifically, the trained Gumpotz model includes: the autonomous vehicle obtains the values of the driving speed of the autonomous vehicle, the driving distance of the midpoint of the rear wheels, and the total length of the available parking space; the The self-driving car uses the driving speed of the self-driving car, the driving distance of the midpoint of the rear wheel, and the total length of the available parking space to train the initial Gangpots model to obtain the trained Gunpots model.

进一步地，所述自动驾驶汽车利用所述当前时间的轨迹曲线，计算待移动的曲线距离值包括：所述自动驾驶汽车根据所述当前时间的轨迹曲线，计算所述轨迹曲线的弧长值；所述自动驾驶汽车根据所述轨迹曲线的弧长值，计算待移动的曲线距离。Further, calculating the distance value of the curve to be moved by the self-driving car using the trajectory curve of the current time includes: the self-driving car calculates the arc length value of the trajectory curve according to the trajectory curve of the current time; The self-driving car calculates the curve distance to be moved according to the arc length value of the trajectory curve.

进一步地，所述自动驾驶汽车利用所述当前时间的轨迹曲线，分别计算待移动的角度包括：所述自动驾驶汽车根据当前时间的距离值和所述轨迹曲线的弧长值，构造所述自动驾驶汽车当前时间的可用停车空间的平均总长度等式；所述自动驾驶汽车将所述可用停车空间的平均总长度等式和时间输入到所述初始冈波茨模型中，得到多个不同的冈波茨模型值；所述自动驾驶汽车根据所述多个不同的冈波茨模型值，计算待移动的角度。Further, the self-driving car uses the trajectory curve of the current time to calculate the angle to be moved respectively, including: the self-driving car constructs the automatic driving vehicle according to the distance value of the current time and the arc length value of the trajectory curve. The average total length equation of the available parking space at the current time of the driving car; the self-driving car inputs the average total length equation and time of the available parking space into the initial Gunpotz model to obtain a number of different Gunpots model value; the autonomous vehicle calculates the angle to be moved according to the plurality of different Gunpots model values.

进一步地，所述自动驾驶汽车根据所述曲线距离值和角度进行精准停车包括：所述自动驾驶汽车根据所述曲线距离值和角度进行移动停车处理，得到移动停车结果；所述自动驾驶汽车判断所述移动停车结果是否已达到精准停车；若判断移动停车结果未达到精准停车，则所述自动驾驶汽车重新计算新的曲线距离值和新的角度，直至达到精准停车。Further, the precise parking of the self-driving car according to the curve distance value and the angle includes: the self-driving car performs a mobile parking process according to the curve distance value and angle to obtain a mobile parking result; the self-driving car determines Whether the mobile parking result has reached precise parking; if it is determined that the mobile parking result has not reached precise parking, the self-driving car recalculates a new curve distance value and a new angle until accurate parking is achieved.

图2是本发明实施例提供的一种自动驾驶汽车精准停车控制装置的示意图，如图2所示，包括：获取模块201，用于在自动驾驶汽车停车期间，分别获取当前时间的距离值、角速度值以及加速度值；输入模块202，用于将所述距离值、角速度值以及加速度值输入到已训练好的冈波茨模型中，得到当前时间的轨迹曲线；计算及停车模块203，用于利用所述当前时间的轨迹曲线，分别计算待移动的曲线距离值和角度，以便根据所述曲线距离值和角度进行精准停车。FIG. 2 is a schematic diagram of a precise parking control device for an autonomous vehicle provided by an embodiment of the present invention. As shown in FIG. 2 , it includes: anacquisition module 201 for acquiring the distance value of the current time, The angular velocity value and the acceleration value; theinput module 202 is used to input the distance value, the angular velocity value and the acceleration value into the trained Gangpotz model to obtain the trajectory curve of the current time; the calculation andparking module 203 is used for Using the trajectory curve at the current time, the distance value and angle of the curve to be moved are calculated respectively, so as to perform precise parking according to the curve distance value and angle.

其中，所述距离值包括以下任一或组合：前左车轮距离前障碍物或前停车线的第一距离值；前右车轮距离前障碍物或前停车线的第二距离值；后左车轮距离后障碍物或后停车线的第三距离值；后右车轮距离后障碍物或后停车线的第四距离值。Wherein, the distance value includes any one or a combination of the following: the first distance value of the front left wheel from the front obstacle or the front stop line; the second distance value of the front right wheel from the front obstacle or the front stop line; the rear left wheel The third distance value from the rear obstacle or the rear stop line; the fourth distance value of the rear right wheel from the rear obstacle or the rear stop line.

进一步地，所述获取模块201包括：通过摄像头和激光距离扫描仪测量当前时间的距离值；通过单轴陀螺仪传感器获取当前时间的角速度值；通过三维加速度传感器测量当前时间的加速度值。Further, the obtainingmodule 201 includes: measuring the distance value at the current time through a camera and a laser distance scanner; obtaining the angular velocity value at the current time through a single-axis gyro sensor; and measuring the acceleration value at the current time through a three-dimensional acceleration sensor.

本发明实施例为了让自动驾驶汽车停车更加快速和精准，在自动驾驶汽车停车路径规划阶段，利用优化方案实时确定轨迹参数，以生成候选的停车路径，主要包括以下步骤：In the embodiment of the present invention, in order to make the parking of the self-driving car more rapid and accurate, in the stage of planning the parking path of the self-driving car, the optimization scheme is used to determine the trajectory parameters in real time to generate a candidate parking path, which mainly includes the following steps:

步骤一， 假设自动驾驶汽车的最大旋转角度为180°，主板控制器的型号为AMD_Geode_LX800，传感器主要使用基于CCD芯片组的VGA接口一体摄像机、单轴陀螺仪传感器、轮轴编码器、近距离传感器、三维加速度传感器，激光距离扫描仪。Step 1: Assume that the maximum rotation angle of the self-driving car is 180°, the model of the motherboard controller is AMD_Geode_LX800, and the sensors mainly use VGA interface integrated cameras based on CCD chipsets, single-axis gyro sensors, wheel encoders, proximity sensors, 3D accelerometer, laser distance scanner.

步骤二，读取传感器测量的自动驾驶汽车此刻的驾驶速度为V，同时读取编码器的刻度数据可以得出自动驾驶汽车此时后轮1和后轮2的行驶距离分别为L₁和L₂，两后轮的中点表示为m，可以算出后轮中点m₂的行驶距离为

。Step 2: Read the driving speed of the self-driving car measured by the sensor as V at the moment, and read the scale data of the encoder at the same time. It can be concluded that the driving distance of the rear wheel 1 and the rear wheel 2 of the self-driving car at this time are L₁ and L respectively.₂ , the midpoint of the_two rear wheels is expressed as m, and the driving distance of the midpoint m2 of the rear wheels can be calculated as

.

步骤三，将此刻的重点m作为输入系统中作为当前车辆位置的一个参考点。引入Gompertz模型的公式

，其中Y为在某个时间用Gompertz预测出来停车轨迹曲线，K代表轨迹宽度，即第一次车辆平行时与车辆停泊完成时，两条平行线之间的距离，用于定义

的上渐进线，e代表一个自然常数，a代表平行停车时，车辆的尾部开始平行于路边线的距离，b表示当前可用停车空间的总长度，t表示车辆行驶的时间变量。Step 3, take the key point m at the moment as a reference point in the input system as the current vehicle position. Introducing the formulation of the Gompertz model

, where Y is the parking trajectory curve predicted by Gompertz at a certain time, and K represents the trajectory width, that is, the distance between the two parallel lines when the vehicle is parallel for the first time and when the vehicle is parked, which is used to define

The upper asymptote of , e represents a natural constant, a represents the distance from the rear of the vehicle parallel to the curb line when parallel parking, b represents the total length of the currently available parking space, and t represents the time variable of the vehicle traveling.

步骤四，带入传感器所获取的可用停车空间的总长度值到Gompertz模型的公式

中，算出一条初步的行动曲线L。Step 4: Bring the total length of the available parking space obtained by the sensor into the formula of the Gompertz model

, calculate a preliminary action curve L.

步骤五，让t趋向于-∞，并将t的值带入步骤三的 Gompertz模型公式

中，可以得出等式如下：Step 5, let t tend to -∞, and bring the value of t into the Gompertz model formula of step 3

, the following equation can be obtained:

；

;

步骤六，利用步骤四所列等式求出的值可以算出Y=K，因此可以利用K的值作为

的上渐进线也就是整个停车场的宽度Step 6, Y=K can be calculated by using the value obtained from the equation listed in Step 4, so the value of K can be used as

The upper asymptote is the width of the entire parking lot

步骤七，将传感器所接收到的相应值带入

中，可以预算出当前时间的轨迹曲线Y_t，Step 7: Bring the corresponding value received by the sensor into

, the trajectory curve Y_t of the current time can be estimated,

其中，传感器所接收到的相应值包括：单轴陀螺仪传感器接收的角速度，近距离传感器测量与障碍物的距离，以及摄像头和激光距离扫描仪测量离停车线的距离，三维加速度传感器测量的此时汽车的加速度值。其中通过摄像头和激光距离扫描仪测量出的距离得到轨迹宽度K值以及可用停车长度b,通过汽车的加速度和角速度计算停车时汽车平移距离，即可得到a的值。Among them, the corresponding values received by the sensor include: the angular velocity received by the single-axis gyro sensor, the distance from the obstacle measured by the proximity sensor, the distance from the parking line measured by the camera and the laser distance scanner, and the distance measured by the three-dimensional acceleration sensor. the acceleration value of the car. Among them, the track width K value and the available parking length b are obtained by the distance measured by the camera and the laser distance scanner, and the value of a can be obtained by calculating the translation distance of the car during parking by the acceleration and angular velocity of the car.

步骤八，再利用函数的弧长公式

，可以初步估算出曲线Y_t的弧长值S，S也初步表示下一步汽车所需要的移动的曲线距离。Step 8: Reuse the arc length formula of the function

, the arc length value S of the curve Y_t can be preliminarily estimated, and S also preliminarily represents the curve distance that the car needs to move in the next step.

步骤九，再算出下一步汽车所需要的移动的曲线距离值之后，我们需要算出下一步汽车需要转向的角度

。先连接后轮中点m₂和前轮中点m₁生成一条中线M，根据常识可以得出M和车身边缘几乎平行。Step 9, after calculating the curve distance value that the car needs to move in the next step, we need to calculate the angle that the car needs to turn in the next step

. First connect the mid-point m₂ of the rear wheel and the mid-point m₁ of the front wheel to generate a center line M. According to common sense, it can be concluded that M is almost parallel to the edge of the vehicle body.

步骤十，利用传感器读取当前时间，自动驾驶汽车前左轮与前右轮与前标线或后标线的距离，后左轮与后右轮与后标线或后车的距离。Step ten, use the sensor to read the current time, the distance between the front left wheel and the front right wheel of the self-driving car and the front marking line or the rear marking line, and the distance between the rear left wheel and the rear right wheel and the rear marking line or the rear car.

步骤十一：把前左轮与前标线或车的距离记为x₁，前右轮与前标线或车的距离记为x₂，后左轮与后标线或车的距离记为x₃，后右轮与后标线或车的距离记为x₄。Step 11: Mark the distance between the front left wheel and the front marking line or car as x₁ , the distance between the front right wheel and the front marking line or car as x₂ , and the distance between the rear left wheel and the rear marking line or car as x₃ , the distance between the rear right wheel and the rear marking or car is recorded as x₄ .

步骤十二：根据上一步所读取出的距离值以及弧长值S我们可以构造一个一元三次方程表示当前时间可用停车空间的平均总长度

，其中

分别为正整常数。Step 12: According to the distance value and arc length value S read in the previous step, we can construct a one-dimensional cubic equation to represent the average total length of the available parking space at the current time

,in

are positive integer constants, respectively.

步骤十三：将算出的当前可用停车空间平均总长度b₁与时间t相乘可以得出等式：Step 13: Multiply the calculated average total length b₁ of the currently available parking space by the time t to obtain the equation:

步骤十四：将步骤十三所得出的b×t带入步骤三的Gompertz模型的公式

中，可以得出等式：Step 14: Bring the b×t obtained in Step 13 into the formula of the Gompertz model in Step 3

, the equation can be derived:

步骤十五：将1~x₁分别带入

中，将1~x₂分别带入

中，将1~x₃分别带入

中，将1~x₄分别带入

中，每次增加1；记

的最大值为max；Step 15: Bring 1~x₁ into

, bring 1~x₂ into

, bring 1~x₃ into

, bring 1~x₄ into

, increment by 1 each time;

The maximum value is max;

步骤十六：重复步骤十三和十四，可以算出max个不同的Y值，分别记为

；Step 16: Repeat steps 13 and 14 to calculate max different Y values, which are recorded as

;

步骤十七：算出

的平均值，记为

Step Seventeen: Calculate

the average value of , denoted as

步骤十八：根据上一步求出的平均值和

的值，可以计算出自动驾驶汽车停车转向的角度

；其中，

Step 18: According to the average value obtained in the previous step and

The value of , can calculate the angle at which the self-driving car stops and turns

;in,

步骤十八，计算机传送角度

值以及曲线轨迹的距离值S到控制器，则可以控制计算器下一步所需要的移动。Step 18, the computer transmits the angle

value and the distance value S of the curve trajectory to the controller, you can control the movement required by the calculator for the next step.

步骤十九，重复步骤四到十八，直到自动驾驶汽车准确停车为止。Step nineteen, repeat steps four to eighteen until the self-driving car stops accurately.

根据本发明实施例提供的方案，具有以下有益效果：The solution provided according to the embodiment of the present invention has the following beneficial effects:

1）考虑了物理上可以实现的最大转向角度，极大的适应了实际生活中自动驾驶汽车的转向情况。1) Considering the maximum steering angle that can be achieved physically, it greatly adapts to the steering situation of autonomous vehicles in real life.

2）在生成候选轨迹的时候对轨迹环境下了碰撞预测检测，极大的保证了自动驾驶汽车的停车安全性。2) When the candidate trajectory is generated, the collision prediction and detection are performed on the trajectory environment, which greatly ensures the parking safety of the self-driving car.

3）利用三次插值法将行驶轨迹参数化为弧长格式，并根据数据多次模拟计算，选出最优的行驶轨迹。3) Use the cubic interpolation method to parameterize the driving trajectory into an arc-length format, and select the optimal driving trajectory according to the data for multiple simulation calculations.

4）不需要进行过多复杂的运算，从而进一步节省了控制时间。4) There is no need to perform too many complicated operations, thereby further saving the control time.

尽管上文对本发明进行了详细说明，但是本发明不限于此，本技术领域技术人员可以根据本发明的原理进行各种修改。因此，凡按照本发明原理所作的修改，都应当理解为落入本发明的保护范围。Although the present invention has been described in detail above, the present invention is not limited thereto, and various modifications can be made by those skilled in the art in accordance with the principles of the present invention. Therefore, all modifications made in accordance with the principles of the present invention should be understood as falling within the protection scope of the present invention.

Claims (10)

1. An automatic automobile accurate parking control method is characterized by comprising the following steps:

during the parking of the automatic driving automobile, the automatic driving automobile respectively acquires a distance value, an angular velocity value and an acceleration value of the current time;

inputting the distance value, the angular velocity value and the acceleration value into a trained Gongworth model by the automatic driving automobile to obtain a track curve of the current time;

and the automatic driving automobile respectively calculates the distance value and the angle of the curve to be moved by utilizing the track curve of the current time so as to accurately stop according to the distance value and the angle of the curve.

2. The precision parking control method of an autonomous automobile according to claim 1, wherein the distance value includes any one or a combination of:

a first distance value of a front left wheel from a front obstacle or a front stop line;

a second distance value of the front right wheel from a front obstacle or a front stop line;

a third distance value of the rear left wheel from a rear obstacle or a rear stop line;

a fourth distance value of the rear right wheel from a rear obstacle or rear stop line.

3. The precision parking control method of an autonomous vehicle as claimed in claim 2, wherein the acquiring the distance value, the angular velocity value and the acceleration value of the current time by the autonomous vehicle comprises:

measuring a distance value of the current time through a camera of the automatic driving automobile and a laser distance scanner;

acquiring an angular speed value of the current time through a single-axis gyroscope sensor of the automatic driving automobile;

and measuring the acceleration value of the current time through the three-dimensional acceleration sensor of the automatic driving automobile.

4. The autonomous precise parking control method of claim 2, wherein the trained Gongborz model comprises:

the automatic driving automobile respectively obtains the driving speed of the automatic driving automobile, the driving distance of the middle point of the rear wheel and the total length value of the available parking space;

and the automatic driving automobile trains the initial Gangbort model by utilizing the running speed of the automatic driving automobile, the running distance of the middle point of the rear wheel and the total length value of the available parking space to obtain the trained Gangbort model.

5. The precision parking control method of an autonomous automobile according to claim 4, wherein the autonomous automobile uses the trajectory curve of the current time, and the calculating of the curve distance value to be moved comprises:

the automatic driving automobile calculates an arc length value of the track curve according to the track curve of the current time;

and the automatic driving automobile calculates the distance of the curve to be moved according to the arc length value of the track curve.

6. The precision parking control method of an autonomous automobile according to claim 5, wherein the autonomous automobile respectively calculating the angles to be moved using the trajectory curve of the current time includes:

the automatic driving automobile constructs an average total length equation of available parking spaces of the automatic driving automobile at the current time according to the distance value of the current time and the arc length value of the track curve;

inputting an average total length equation and time of the available parking space into the initial Golay model by the automatic driving automobile to obtain a plurality of different Golay model values;

and the automatic driving automobile calculates the angle to be moved according to the different Golay model values.

7. The precision parking control method for the automatic driving automobile according to claim 6, wherein the precision parking of the automatic driving automobile according to the curve distance value and the angle comprises the following steps:

the automatic driving automobile carries out mobile parking processing according to the curve distance value and the angle to obtain a mobile parking result;

the automatic driving automobile judges whether the mobile parking result reaches accurate parking;

and if the mobile parking result does not reach the accurate parking result, recalculating a new curve distance value and a new angle by the automatic driving automobile until the accurate parking is achieved.

8. The utility model provides an automatic accurate parking control device of driving car which characterized in that includes:

the acquisition module is used for respectively acquiring a distance value, an angular velocity value and an acceleration value of the current time during the parking of the automatic driving automobile;

the input module is used for inputting the distance value, the angular velocity value and the acceleration value into a trained Gongoni model to obtain a track curve of the current time;

and the calculation and parking module is used for respectively calculating a curve distance value and an angle to be moved by utilizing the track curve of the current time so as to carry out accurate parking according to the curve distance value and the angle.

9. The precision parking control apparatus of an autonomous vehicle of claim 8, wherein the distance value comprises any one or a combination of:

a first distance value of a front left wheel from a front obstacle or a front stop line;

a second distance value of the front right wheel from a front obstacle or a front stop line;

a third distance value of the rear left wheel from a rear obstacle or a rear stop line;

a fourth distance value of the rear right wheel from a rear obstacle or rear stop line.

10. The autonomous driving precision parking control apparatus of claim 9, wherein the obtaining module comprises:

measuring a distance value of the current time through a camera and a laser distance scanner;

acquiring an angular speed value of current time through a single-axis gyroscope sensor;

and measuring the acceleration value of the current time through the three-dimensional acceleration sensor.

Images