Detailed Description
Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Referring to fig. 1, fig. 1 is a flowchart of a fault detection method of an automatic driving vehicle according to an embodiment of the disclosure, where the fault detection method of the automatic driving vehicle includes the following steps:
Step S101, under the condition that a target abnormality occurs in the automatic driving vehicle in the process of performing virtual simulation test on the automatic driving vehicle, determining candidate faults corresponding to the target abnormality in a preset data set, wherein the preset data set comprises candidate faults corresponding to at least two preset abnormalities, the target abnormality is any one of the at least two preset abnormalities, the candidate faults corresponding to the target abnormality are part of the candidate faults contained in the preset data set, and the candidate faults are faults of a preset functional module in the automatic driving vehicle;
Step S102, determining a target fault in candidate faults corresponding to the target abnormality, wherein the target fault is a fault of the preset functional module in the virtual simulation test process.
The virtual simulation test for the automatic driving vehicle can specifically refer to the virtual simulation test for the automatic driving vehicle under various test scenes. The test scene may include various driving scenes that may be encountered during the running process of the automatic driving vehicle, for example, the test scene may include driving scenes such as a pedestrian crossing road scene, a front vehicle lane changing scene, a normal following scene, and the like. By performing simulation test on the automatic driving vehicle based on the test scene, whether corresponding functional modules in the automatic driving vehicle can accurately identify the test scene or not can be determined, and corresponding responses can be made for different test scenes. For example, when virtual simulation testing is performed on the autonomous vehicle based on the pedestrian crossing road scene, it is determined whether the autonomous vehicle is capable of decelerating or reasonably avoiding so that subsequent autonomous vehicles can respond accordingly for various driving scenes.
Specifically, various traffic elements and traffic parameter persons in the test scene can be modeled in advance based on simulation software to obtain a virtual simulation world. The modeling can be performed according to an actual road traffic scene, and an actual attribute parameter is given to each traffic element and traffic parameter value. In addition, it is also possible to fictitious road traffic scenes and assign actual attribute parameters to each traffic element and traffic parameter value.
It may be appreciated that after the modeling process is completed, corresponding attribute information may be assigned to each element in the built virtual simulation world based on preset simulation parameters, where the attribute information includes a position, a speed, a forward direction, status information of each traffic parameter value, and the like. Then, virtual simulation is performed.
The preset data set may include the at least two preset anomalies and candidate faults corresponding to the preset anomalies, where the preset anomalies and the candidate faults corresponding to the preset anomalies may be stored in association. In this way, when it is determined that the autonomous vehicle has a target abnormality, a candidate fault corresponding to the target abnormality may be determined in the preset data set based on a correspondence between a preset abnormality and the candidate fault.
The preset function module may be various function modules in an autonomous vehicle, and for example, may include a sensing module, a positioning module, a map module, and the like.
The preset abnormality may be driving behavior of various abnormalities in the running process of the automatic driving vehicle, for example, the preset abnormality may include abnormality such as reverse driving, red light running, sudden braking, main collision, vehicle shaking, dragon drawing, exiting from an automatic driving mode, and the like. Accordingly, the candidate faults can comprise faults possibly occurring in each functional module in the automatic driving vehicle, for example, the candidate faults can comprise frame loss faults and error faults in the process of detecting the target object by the sensing module, positioning drift faults of the positioning module and the like.
It is understood that the candidate faults corresponding to the preset abnormality may include all faults that can cause the autonomous vehicle to experience the preset abnormality. For example, when the preset abnormality is a primary collision fault, for the perception module, the candidate faults corresponding to the primary collision fault may include faults such as target frame loss, target delay selection, excessive target position error, target type error, and target being a virtual target. The target frame loss is a frame loss fault in the process of detecting the target object. The target delay is selected as the target locking time in the process of detecting the target object is too late, the target position error is too large as the positioning error of the target object, the target type error is the type identification error of the target object, the target is the virtual target, namely the target which is not existed is identified as the entity target, for example, the reflection of the tree is identified as the vehicle.
It is understood that whether the above-described preset abnormality occurs in the autonomous vehicle may be detected in the course of the virtual simulation test based on the detection means in the related art, and the preset abnormality occurring in the autonomous vehicle may be determined as the target abnormality. For example, the occurrence of a reverse failure of the autonomous vehicle may be determined when it is detected that the travel direction of the autonomous vehicle is opposite to the travel direction indicated by the lane in which the autonomous vehicle is located by detecting the travel direction of the autonomous vehicle and comparing the travel direction of the autonomous vehicle with the travel direction indicated by the lane in which the autonomous vehicle is located. In addition, because specific time of the red light of the traffic indicator lights at various positions in the virtual simulation world can be directly obtained from simulation information, and time of the automatic driving vehicle actually entering and exiting a certain traffic intersection can be obtained from detection data of the automatic driving vehicle, when the time of the automatic driving vehicle entering and exiting a certain traffic intersection is detected to be within the red light-on time period of the traffic intersection, the automatic driving vehicle is determined to have red light running abnormality. And when the coincidence of the automatic driving vehicle and the front vehicle is detected in the virtual simulation picture, determining that the collision abnormality occurs. Meanwhile, whether the automatic driving vehicle has shaking faults or picture dragon faults or not can be determined by detecting the positions of the automatic driving vehicle in a plurality of continuous frames in the simulation picture, and when the change values of the transverse positions of the automatic driving vehicle between adjacent pictures exceed a set threshold value in the continuous multi-frame picture. When the automatic driving vehicle exits the automatic driving mode, a corresponding signal for exiting the automatic driving mode is triggered, and when the signal for exiting the automatic driving mode is acquired, the automatic driving vehicle is determined to have a fault for exiting the automatic driving mode.
Accordingly, in the case of determining the candidate fault corresponding to the target abnormality, whether the candidate fault occurs in the preset functional module in the autonomous vehicle or not may be detected based on the detection means in the related art in the process of the virtual simulation test, and the candidate fault output by the autonomous vehicle may be determined as the target fault.
In this embodiment, by establishing a correspondence between a preset abnormality and a candidate fault in advance, when a target abnormality of an autonomous vehicle is detected, the candidate fault corresponding to the target fault may be determined first, and then, the target fault is determined in the target fault, so as to implement a process of completing fault detection. In the embodiment of the disclosure, since the troubleshooting is only performed in the candidate faults corresponding to the target faults, and the candidate faults corresponding to the target faults are only part of the candidate faults included in the preset data set, the troubleshooting range can be effectively reduced, and the fault detection efficiency is further improved.
Optionally, the candidate faults corresponding to the target abnormality include at least two first faults, and the determining the target fault in the candidate faults corresponding to the target abnormality includes:
Acquiring a detection parameter set and a simulation parameter set, wherein the detection parameter set comprises detection attribute parameters for detecting a target object by the automatic driving vehicle in the virtual simulation test process, the simulation parameter set comprises true value attribute parameters of the target object in the virtual simulation test process, and the target object comprises at least one of the automatic driving vehicle and driving environment objects outside the automatic driving vehicle;
the target fault is determined in the at least two first faults based on the detection parameter set and the simulation parameter set.
The first fault may be a fault that may occur in various functional modules in the autonomous vehicle. For example, the method can comprise the steps of detecting frame loss faults, sensing error faults and the like in the process of detecting the target object by the sensing module, positioning drift faults of the positioning module and the like.
Because the automatic driving vehicle generally senses external driving environment information through the sensing module in the process of performing virtual simulation test on the automatic driving vehicle, and then outputs corresponding driving actions based on the sensed external driving environment information. However, when an error occurs in the sensed information, the above-described preset abnormality may be caused to occur in the automatically driven vehicle. And typically, specific detection data errors may result in specific abnormal driving behavior, for example, when detecting errors for a front vehicle, collision of an autonomous vehicle with the front vehicle may result.
Therefore, in the embodiment of the disclosure, when it is required to determine whether a certain first fault occurs in the process of performing the virtual simulation test on the automatically driven vehicle, it is only required to determine whether the detected data of the first fault is abnormal in the process of performing the virtual simulation test on the automatically driven vehicle, and it is able to determine whether the first fault occurs in the automatically driven vehicle in the process of performing the virtual simulation test. Specifically, when the detection data capable of causing the first failure is abnormal, it is determined that the first failure occurs in the automatically driven vehicle in the course of the virtual simulation test.
Based on the above, in the process of performing virtual simulation test on the automatic driving vehicle, information perceived by each module in the automatic driving vehicle is stored in the detection parameter set, so that when determining whether a certain first fault occurs in the automatic driving vehicle, data corresponding to the first fault in the detection parameter set is detected to determine whether the first fault occurs. The information perceived by each module may include original information collected by each module, or may include information after each module processes the original information, that is, the detection parameter set may include original information collected by each module, and information after each module processes the original information.
The detection parameter set includes detection attribute parameters of the automatic driving vehicle and detection attribute parameters of driving environment objects outside the automatic driving vehicle, because the automatic driving vehicle is required to sense the state of the automatic driving vehicle and the driving environment in the virtual simulation test. The detection attribute parameter is an attribute value detected by detecting the target object. For example, when the target object is a vehicle in front of an autonomous vehicle, the detected attribute parameters may include attribute information such as a position, a speed, and the like of the vehicle in front.
Further, since the detection parameter set includes detection attribute parameters for a target object, in order to determine whether data in the detection parameter set is abnormal, it is necessary to acquire real attribute parameters of the target object. In this way, the parameters in the detection parameter set can be compared with the corresponding real attribute parameters in the simulation parameter set to determine whether the detection parameters are correct.
Specifically, during the process of performing the virtual simulation test, the real attribute of each target object in the virtual world at the initial time of the simulation is usually manually input by related personnel, so that the real attribute of the target object can be directly obtained from the simulation. In this way, the set of simulation parameters may be derived from the real properties of the target object given from the simulation software.
In this embodiment, when it is required to determine whether a certain first abnormality occurs in the automated driving vehicle during the virtual simulation test, it may be determined whether the first abnormality occurs by comparing the detected attribute parameter corresponding to the first abnormality with the corresponding real attribute parameter, so as to implement a determination process of the target abnormality.
Optionally, the determining a target fault in the at least two first faults based on the detection parameter set and the simulation parameter set includes:
under the condition that the first attribute parameters are not matched with the second attribute parameters, determining that the first candidate fault is the target fault;
The first candidate faults are faults in the at least two first faults, the first candidate faults are used for representing target attribute abnormality of the preset functional module, the first attribute parameters are attribute parameters in the detection parameter set and used for representing the target attribute of the preset functional module, and the second attribute parameters are attribute parameters in the simulation parameter set and used for representing the target attribute of the preset functional module.
The first candidate fault may be various types of faults that may occur to various functional modules in the autonomous vehicle. For example, the first candidate fault may be any fault in the candidate faults corresponding to the sensing module, that is, the first candidate fault may be any fault in the "target frame loss, target delay selection, target position error is too large, target type error and target is a virtual target".
The foregoing mismatch between the first attribute parameter and the second attribute parameter may mean that the first attribute parameter and the second attribute parameter are not equal, or that a difference between the first attribute parameter and the second attribute parameter exceeds a preset threshold.
In one embodiment of the present disclosure, the test scenario may be a scenario in which a front vehicle changes lanes to a main vehicle lane, where the main vehicle is the above-mentioned autonomous vehicle. The first candidate fault may be target delay selection, the preset functional module is a sensing module, and the target attribute is time for the sensing module to select the target object. When the current car changes lanes and the selected time of the main car to the front car is too late, the main car and the front car can be possibly caused to collide, or the main car is in emergency braking and other anomalies, so that the target anomaly can be a main duty collision or emergency braking.
Specifically, when the front vehicle starts lane change, the main vehicle will select the front vehicle and identify the front vehicle. Thus, in order to determine whether the host vehicle has a delayed selection failure, the first attribute parameter may be a point in time when the perception module of the host vehicle starts selecting the preceding vehicle, and the first attribute parameter may be directly obtained from the detected attribute parameter. And the second attribute parameter may be a point in time when the preceding vehicle begins to change track, which point in time may be directly obtained from the simulation parameter set. In this way, a judgment threshold value can be preset, when the time difference between the first attribute parameter and the second attribute parameter is larger than the judgment threshold value, the delayed selected fault of the host vehicle is determined, and the delayed selected fault is determined as the target fault to be output.
In this embodiment, when it is determined whether an abnormality occurs in a preset function module of an autonomous vehicle, since the abnormality specifically indicates that an abnormality occurs in a target attribute of a specific function module in the autonomous vehicle, a first attribute parameter corresponding to the target attribute may be obtained from a detection parameter set, and a second attribute parameter corresponding to the target attribute may be obtained from a simulation parameter set, and then, it is determined whether the first attribute parameter and the second attribute parameter are matched to determine whether the abnormality occurs in the autonomous vehicle in a virtual simulation test.
Optionally, the preset functional module includes a sensing module and a positioning module, and before the determining that the first candidate fault is the target fault, the method further includes:
Determining a matching state of the first attribute parameter and the second attribute parameter based on a first matching condition under the condition that the first candidate fault is a fault of the perception module; determining a matching state of the first attribute parameter and the second attribute parameter based on a second matching condition under the condition that the first candidate fault is a fault of the positioning module;
wherein the first matching condition is different from the second matching condition.
The first matching condition may include a judging condition for judging whether each attribute parameter in the sensing module is abnormal, and correspondingly, the second matching condition may include a judging condition for judging whether each attribute parameter in the positioning module is abnormal. And the first matching condition and the second matching condition may be preset.
It will be appreciated that the target anomalies may be anomalies resulting from faults in the sensing module, as well as in the locating module. In addition, the fault of the sensing module and the fault of the positioning module may correspond to the same preset abnormality. Thus, the at least two first faults may comprise a fault of the perception module and/or a fault of the localization module.
In this embodiment, by presetting the first matching condition and the second matching condition, when it is required to determine whether a corresponding first fault occurs in the sensing module and the positioning module in the virtual simulation test process of the automatic driving vehicle, the determination process of the target fault may be implemented based on the first matching condition and the second matching condition.
Optionally, the first matching condition includes at least one of:
Determining that the first attribute parameter is not matched with the second attribute parameter when the first candidate fault is a frame loss fault, the detection frame number of the target object is not matched with the shooting frame number of the target object, the first attribute parameter comprises the detection frame number, and the second attribute parameter comprises the shooting frame number;
Determining that the first attribute parameter is not matched with the second attribute parameter when the first candidate fault is a delayed selected fault and a difference between a first time point and a second time point is greater than a first threshold, wherein the first time point is an actual selected time point of the automatic driving vehicle on the target object, the second time point is a time point when the target object enters a detection range of the automatic driving vehicle, the first attribute parameter comprises the first time point, and the second attribute parameter comprises the second time point;
determining that the first attribute parameter does not match the second attribute parameter in the case that the first candidate fault is a detection fault of an attribute of a target obstacle and the first attribute detection parameter of the target obstacle does not match a first attribute real parameter of the target obstacle, the target object including the target obstacle, the first attribute parameter including the first attribute detection parameter, the second attribute parameter including the first attribute real parameter, the attribute of the target obstacle including at least one of a position attribute, a speed attribute, and a type attribute;
And under the condition that the first candidate fault is a detection fault of an attribute of a lane line and a second attribute detection parameter of the lane line is not matched with a second attribute real parameter of the lane line, determining that the first attribute parameter is not matched with the second attribute parameter, wherein the target object comprises the lane line, the first attribute parameter comprises the second attribute detection parameter, the second attribute parameter comprises the second attribute real parameter, and the attribute of the lane line comprises at least one of a morphological attribute of the lane line and a position attribute of the lane line.
The target object may be a preset object, for example, in a front vehicle lane change test scenario, the target object may be a front vehicle. The method provided by the embodiment of the disclosure is further explained by taking the target object as a front vehicle in a front vehicle lane change testing scene as an example.
The frame loss detection fault may be a fault that a missed detection exists in the process of identifying the target object. Specifically, after the host vehicle selects the target object, the target object is continuously photographed, and the photographed image is transmitted to the sensing module, and the sensing module recognizes the target object in the photographed image to determine the real-time position of the target object. However, in the process of identifying the target object in the captured image by the sensing module, there may be a problem of missed detection, that is, the image content of the target object is included in a certain frame of captured image, but the sensing module does not output the position of the target object for the frame of captured image, so that an incorrect driving decision may be caused for the automatic driving vehicle. The number of detection frames of the target object may be the number of shot images including the target object detected by the sensing module in the virtual simulation test process, where the number may be directly obtained from the detection parameter set. The number of times of shooting the target object after the host vehicle selects the target object can be directly obtained from the simulation parameter set, so that whether the frame loss fault occurs can be determined by determining whether the detection frame number of the target object is equal to the shooting frame number of the target object.
The delay selected fault is a fault which is selected for the target object and is too late in time. The above detection range may refer to a lane where the host vehicle is located, that is, the second time point is a time actual point when the target object starts lane change. Therefore, the delay of the target object selected by the host vehicle can be determined by comparing the first time point of the target object selected by the host vehicle with the time point of the actual track change of the target object, and when the delay is larger than a first threshold value, the delayed selection fault of the host vehicle is determined.
The detection failure of the attribute of the target obstacle may specifically include a position detection failure, a speed detection failure, and a type detection failure. The above-described position detection failure may mean that an error between the detected position of the target obstacle and the actual position of the target obstacle exceeds a preset position error. The above-described speed detection failure may mean that an error between the detected speed of the target obstacle and the actual speed of the target obstacle exceeds a preset speed error. The above-described type of detection failure is a detection error of the type of the target obstacle, for example, a pedestrian is erroneously detected as a vehicle. The first attribute detection parameters may include a position detection parameter, a speed detection parameter, and a type detection parameter for the target obstacle. The first attribute real parameters may include a real position parameter, a real speed parameter, and a real type parameter of the target obstacle.
Accordingly, when the first candidate fault is a fault for detecting the position of the target obstacle, and the position detection parameter of the target obstacle is not matched with the real position parameter of the target obstacle, the first attribute parameter and the second attribute parameter are determined to be not matched. Wherein the mismatch of the position detection parameter and the real position parameter may mean that a distance between a position indicated by the position detection parameter and a position indicated by the real position parameter exceeds a preset distance threshold.
And when the first candidate fault is a speed detection fault of the target obstacle and the speed detection parameter of the target obstacle is not matched with the real speed parameter of the target obstacle, determining that the first attribute parameter is not matched with the second attribute parameter. Wherein the speed detection parameter not matching the real speed parameter may mean that a distance between a speed indicated by the speed detection parameter and a speed indicated by the real speed parameter exceeds a preset speed threshold.
And when the first candidate fault is a type detection fault of the target obstacle and the type detection parameter is different from the real type parameter, determining that the first attribute parameter is not matched with the second attribute parameter.
The shape attribute of the lane line can comprise the length of the lane line and whether the lane line is missing, and the position attribute of the lane line can comprise a lane line change state attribute, the relative position attribute of two lane lines of the same lane and the position error attribute of the lane line. Correspondingly, the detection faults of the attributes of the lane lines can comprise lane line length detection faults, lane line jitter faults, lane line internal and external eight faults, lane line single and double-side missing faults and lane line position error faults.
Accordingly, the second attribute detection parameters may include a detection length of the lane line, a detected lane line position change state, a detected lane line shape, a detected lane line position. The second attribute real parameters may include a real length of the lane line, a real lane line position, a real lane line shape.
And when the first candidate fault is a lane line length detection fault and the difference value between the detected length of the lane line and the real length of the lane line is smaller than a preset length threshold value, determining that the first attribute parameter is not matched with the second attribute parameter.
And when the first candidate fault is a lane line jitter fault, and the detected lane line position change state indicates that the standard deviation of the position error change quantity between the positions of the lane lines obtained by continuous multi-frame shooting exceeds a preset threshold value, determining that the first attribute parameter is not matched with the second attribute parameter.
And when the first candidate fault is an internal and external eight fault of the lane line and the detected lane line form indicates that the slope of the lane line in the continuous multi-frame images exceeds a preset slope threshold value, determining that the first attribute parameter is not matched with the second attribute parameter.
And when the first candidate fault is a lane line single-double-side missing fault and a plurality of continuous frames of actual lane lines exist in simulation time but the sensing lane line is not detected, determining that the first attribute parameter is not matched with the second attribute parameter.
And when the first candidate fault is a lane line position error fault and the difference between the detected lane line position and the actual lane line position is larger than a set position error threshold value, determining that the first attribute parameter is not matched with the second attribute parameter.
Please refer to the following table 1 for a definition of a fault existing in the sensing module in an embodiment of the present disclosure, please refer to the following table 2, which is a corresponding relationship between the first matching condition and the detection policy corresponding to each fault type:
TABLE 1
TABLE 2
Referring to table 3, table 3 is a candidate fault of the sensing module corresponding to the preset abnormality.
TABLE 3 Table 3
In table 3, "O" indicates that there is a correspondence, and "X" indicates that there is no correspondence, for example, candidate faults in the sensing module corresponding to the preset abnormal "main-responsibility collision" include target frame loss, target delay selection, excessive target speed error, excessive target position error, and target type error.
In this embodiment, by determining the possible fault types of the sensing module in advance and presetting a detection policy corresponding to each fault, when it is required to determine whether the sensing module has a corresponding fault in the virtual simulation test process, the detection may be performed directly by adopting the preset detection policy, so as to implement the fault detection process of the sensing module.
Optionally, the second matching condition includes at least one of:
Determining that the first attribute parameter does not match the second attribute parameter, the first attribute parameter including the detected location parameter, the second attribute parameter including the true location parameter, and the location drift fault including a location drift fault and a heading angle drift fault, if the first candidate fault is a location drift fault of the autonomous vehicle and the detected location parameter of the autonomous vehicle does not match the true location parameter of the autonomous vehicle;
And under the condition that the first candidate fault is a positioning jump fault of the automatic driving vehicle and the first position change parameter is not matched with the second position change parameter, determining that the first attribute parameter is not matched with the second attribute parameter, wherein the first position change parameter is a position change parameter detected by the automatic driving vehicle, the second position change parameter is a real position change parameter of the automatic driving vehicle, the first attribute parameter comprises the first position change parameter, the second attribute parameter comprises the real positioning parameter, and the positioning jump fault comprises a positioning jump fault and a course angle jump fault.
Please refer to the following table 4 for a definition of a fault existing in the positioning module in an embodiment of the present disclosure, please refer to the following table 5, which is a corresponding relationship between the second matching condition and the detection policy corresponding to each fault type:
TABLE 4 Table 4
TABLE 5
The position error described in table 5 refers to the coordinates of the center point of the rear axle of the vehicle output by the positioning module and the course angle error are the same in a cartesian coordinate system established by taking the center of the rear axle of the automatic driving vehicle in simulation as the origin, taking the front of the vehicle as the x axis, taking the side as the y axis and taking the vertical as the z axis.
In addition, according to the situation of the road test problem, the positioning drift may cause the vehicle to shake/draw, collide with risk or collide with road shoulder and other abnormal driving behaviors, and the positioning jump may cause the vehicle to exit from the automatic driving abnormal behavior, so similar to the analysis of the perception problem, the connection between the main vehicle abnormal behavior and the positioning problem can be established, and the traceability analysis of the main vehicle abnormal behavior problem is facilitated.
In this embodiment, by determining the possible fault types of the positioning module in advance and presetting a detection policy corresponding to each fault, when it is required to determine whether the positioning module has a corresponding fault in the virtual simulation test process, the fault detection process of the positioning module can be implemented by directly adopting the preset detection policy.
Optionally, the preset functional module further includes a map module, the candidate faults corresponding to the target abnormality further includes at least two second faults, the second faults are faults of the map module, and the determining the target fault in the candidate faults corresponding to the target abnormality further includes:
and determining the target fault in the at least two second faults under the condition that the first candidate faults are determined to be faults of other modules except the sensing module and the positioning module based on the first matching condition and the second matching condition and no faults exist in the other modules except the map module.
Because the map module generally provides indication for other modules, it does not need to collect data or process data, and therefore, it is unable to detect faults by a method similar to the positioning module and the sensing module, that is, it is unable to detect whether faults exist by comparing the detected value with the true value. In the related art, a map is mainly provided, the failure rate is relatively low, and the failure rates of the sensing module and the positioning module are relatively high. Therefore, when the autonomous vehicle has a target abnormality, the cause of the target abnormality may be the map module in the case where it is determined that there is no failure in both the perception module and the positioning module. And performing fault detection in the candidate faults corresponding to the map module.
And determining that the first candidate fault is a fault of a module other than the sensing module and the positioning module based on the first matching condition and the second matching condition, namely determining that the sensing module and the positioning module have no fault in the virtual simulation test process based on the first matching condition and the second matching condition.
It is understood that the absence of a fault in a module other than the map module means that the map module, the sensing module, and the positioning module do not have a fault in the process of the virtual simulation test. Specifically, to ensure that the map module, the sensing module, and the other modules except the positioning module do not fail during the virtual simulation test, the map module, the sensing module, and the other modules except the positioning module may be set to an ideal state in simulation software. In this way, in the virtual simulation test process, when the target abnormality occurs, only the abnormality may occur in the modules of the map module, the sensing module and the positioning module. And in the case that it is determined that the sensing module and the positioning module have no fault based on the first matching condition and the second matching condition, the target abnormality may be considered to be an abnormality caused by a fault of the map module. Thus, the target fault may be determined among the at least two second faults.
In this embodiment, when the first candidate fault is determined to be a fault of a module other than the sensing module and the positioning module based on the first matching condition and the second matching condition, and the fault of the module other than the map module does not exist, the target fault is determined in the at least two second faults, so that it is ensured that the fault detection can be performed quickly even when the map module fails.
Optionally, the at least two second faults include a second candidate fault, the autonomous vehicle performs the virtual simulation test based on map information in the map module, and determining the target fault in the at least two second faults includes:
Determining the second candidate fault as the target fault under the condition that the second candidate fault is a target map element marking fault and the marking fault exists in the target map element in the map information, wherein the target map element comprises at least one of a speed limit identifier, a lane line, a traffic indicator and a lane;
And under the condition that the second candidate faults are binding relation faults among map elements and the binding relation faults among the map elements in the map information are detected, determining the second candidate faults as the target faults, wherein the binding relation comprises at least one of binding relations among different lanes and binding relations among traffic indicator lamps and lanes.
Referring to the following table 6, possible faults of the map module may include an element miss-labeling fault, an element mislabeling fault, an element binding relationship missing fault, an element binding relationship error fault and an element labeling error oversize fault.
TABLE 6
The map speed limit missing refers to the speed limit mark missing of the lane where the main vehicle is located, and at the moment, overspeed of the main vehicle may be caused. The preceding (succeeding) missing of the host vehicle lane means that the road in front of or behind the lane in which the host vehicle is located is missing due to the map.
Specifically, a correspondence relationship between the candidate fault and the preset abnormality in the map module may be established in advance. In this way, when it is determined that the host vehicle has a target abnormality in the virtual simulation test process and it is determined that the abnormality is not caused by the sensing module and the positioning module, the at least two second faults may be determined from among the faults that may exist in the map module using the correspondence relationship based on the abnormality type according to the target abnormality. And detecting whether the second fault exists in the map module by adopting a means in the related art so as to determine a target fault.
In the virtual simulation test, when a host vehicle runs out or draws, the smoothness of a map lane is insufficient or the lane inheritance relationship is bound incorrectly, when the host vehicle exits from an automatic driving mode, the lane inheritance relationship is lost, when the vehicle runs out, the lane direction is marked incorrectly, and when the vehicle runs out of a red light, the traffic light corresponding to the host vehicle lane is unbound or bound incorrectly.
In this embodiment, by determining the possible fault types of the map module in advance and presetting preset anomalies corresponding to each fault type, when a target anomaly occurs in the host vehicle, at least two second faults can be screened out of the possible faults of the map module, so that the efficiency of fault detection is improved.
Optionally, the preset anomaly includes at least one of a collision, a braking deceleration exceeding a preset deceleration, a vehicle sway exceeding a preset amplitude, exiting an automatic driving mode, reversing, and running a red light.
The collision can refer to main vehicle responsibility collision, sudden braking occurs when the braking deceleration exceeds the preset deceleration, and vehicle shaking amplitude exceeds the preset amplitude, namely vehicle shaking.
It can be understood that means in the related art may be used to detect whether the foregoing preset abnormality occurs in the autonomous vehicle during the virtual simulation test, and when the foregoing abnormality occurs, abnormality information may be output, so that a subsequent relevant person determines a cause of the foregoing preset abnormality based on the fault detection method provided in the present disclosure.
In this embodiment, by presetting the types of the various preset anomalies and setting the candidate faults corresponding to each preset anomaly, when the preset anomalies occur in the process of virtual simulation testing of the automatic driving vehicle, the fault detection range can be reduced, and the fault detection efficiency can be improved.
It should be noted that, when the embodiments of the present disclosure are specifically illustrated, only the sensing module, the map module and the positioning module are described. The methods of the embodiments of the present disclosure may also be applied to fault detection processes of other modules of an autonomous vehicle.
In addition, in another embodiment of the present disclosure, in the process of performing the virtual simulation test, the above fault detection method may be used to test each module in the autonomous vehicle one by one. For example, when testing the map module, other modules outside the map module are set to an ideal state, then the map module is tested based on various test scenes, then, in the case of occurrence of a target fault, a candidate fault in the map module corresponding to the target fault is determined based on the correspondence between the preset abnormality and the candidate fault, and fault detection is performed from the determined candidate faults.
First, because fault detection can be carried out only under the condition that the automatic driving vehicle is abnormal, related personnel can pay attention to the abnormal problem of the automatic driving vehicle at first, the positioning efficiency of the problem is improved, and the situation that the indexes are out of limit but do not influence the operation of the automatic driving vehicle is ignored. Second, since each module is detected only when abnormal driving behavior occurs, the simulation operation efficiency can be improved, and the calculation resources can be saved.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a fault detection device 200 for an automatic driving vehicle according to an embodiment of the disclosure, where the fault detection device 200 for an automatic driving vehicle includes:
A first determining unit 201, configured to determine, in a preset dataset, a candidate fault corresponding to a target abnormality when a target abnormality is detected in the autonomous vehicle during a virtual simulation test of the autonomous vehicle, where the preset dataset includes candidate faults corresponding to at least two preset abnormalities, the target abnormality is any one of the at least two preset abnormalities, and the candidate fault corresponding to the target abnormality is a part of candidate faults included in the preset dataset, and the candidate faults are faults of a preset function module in the autonomous vehicle;
And a second determining unit 202, configured to determine a target fault from candidate faults corresponding to the target abnormality, where the target fault is a fault occurring in the preset functional module during the virtual simulation test.
Optionally, referring to fig. 3, the candidate faults corresponding to the target abnormality include at least two first faults, and the second determining unit 202 includes:
An acquisition subunit 2021 configured to acquire a detection parameter set and a simulation parameter set, where the detection parameter set includes a detection attribute parameter for detecting a target object by the autonomous vehicle during the virtual simulation test, and the simulation parameter set includes a true value attribute parameter of the target object during the virtual simulation test;
A determination subunit 2022 is configured to determine the target fault in the at least two first faults based on the detection parameter set and the simulation parameter set.
Optionally, the determining subunit 2022 is specifically configured to determine, in a case where the first attribute parameter does not match the second attribute parameter, that the first candidate fault is the target fault;
The first candidate faults are faults in the at least two first faults, the first candidate faults are used for representing target attribute abnormality of the preset functional module, the first attribute parameters are attribute parameters in the detection parameter set and used for representing the target attribute of the preset functional module, and the second attribute parameters are attribute parameters in the simulation parameter set and used for representing the target attribute of the preset functional module.
Optionally, the preset functional module includes a sensing module and a positioning module, where the determining subunit 2022 is specifically further configured to determine, based on a first matching condition, a matching state of the first attribute parameter and the second attribute parameter when the first candidate fault is a fault of the sensing module;
wherein the first matching condition is different from the second matching condition.
Optionally, the first matching condition includes at least one of:
Determining that the first attribute parameter is not matched with the second attribute parameter when the first candidate fault is a frame loss fault, the detection frame number of the target object is not matched with the shooting frame number of the target object, the first attribute parameter comprises the detection frame number, and the second attribute parameter comprises the shooting frame number;
Determining that the first attribute parameter is not matched with the second attribute parameter when the first candidate fault is a delayed selected fault and a difference between a first time point and a second time point is greater than a first threshold, wherein the first time point is an actual selected time point of the automatic driving vehicle on the target object, the second time point is a time point when the target object enters a detection range of the automatic driving vehicle, the first attribute parameter comprises the first time point, and the second attribute parameter comprises the second time point;
determining that the first attribute parameter does not match the second attribute parameter in the case that the first candidate fault is a detection fault of an attribute of a target obstacle and the first attribute detection parameter of the target obstacle does not match a first attribute real parameter of the target obstacle, the target object including the target obstacle, the first attribute parameter including the first attribute detection parameter, the second attribute parameter including the first attribute real parameter, the attribute of the target obstacle including at least one of a position attribute, a speed attribute, and a type attribute;
And under the condition that the first candidate fault is a detection fault of an attribute of a lane line and a second attribute detection parameter of the lane line is not matched with a second attribute real parameter of the lane line, determining that the first attribute parameter is not matched with the second attribute parameter, wherein the target object comprises the lane line, the first attribute parameter comprises the second attribute detection parameter, the second attribute parameter comprises the second attribute real parameter, and the attribute of the lane line comprises at least one of a morphological attribute of the lane line and a position attribute of the lane line.
Optionally, the second matching condition includes at least one of:
Determining that the first attribute parameter does not match the second attribute parameter, the first attribute parameter including the detected location parameter, the second attribute parameter including the true location parameter, and the location drift fault including a location drift fault and a heading angle drift fault, if the first candidate fault is a location drift fault of the autonomous vehicle and the detected location parameter of the autonomous vehicle does not match the true location parameter of the autonomous vehicle;
And under the condition that the first candidate fault is a positioning jump fault of the automatic driving vehicle and the first position change parameter is not matched with the second position change parameter, determining that the first attribute parameter is not matched with the second attribute parameter, wherein the first position change parameter is a position change parameter detected by the automatic driving vehicle, the second position change parameter is a real position change parameter of the automatic driving vehicle, the first attribute parameter comprises the first position change parameter, the second attribute parameter comprises the real positioning parameter, and the positioning jump fault comprises a positioning jump fault and a course angle jump fault.
Optionally, the preset functional module further includes a map module, the candidate fault corresponding to the target abnormality further includes at least two second faults, the second faults are faults of the map module, and the determining subunit 2022 is specifically further configured to determine, when the first candidate fault is a fault of another module other than the sensing module and the positioning module based on the first matching condition and the second matching condition, and the other module other than the map module has no fault, determine the target fault in the at least two second faults.
Optionally, the at least two second faults include a second candidate fault, the autonomous vehicle performs the virtual simulation test based on map information in the map module, and the determining subunit 2022 is specifically further configured to determine that the second candidate fault is a target fault if the second candidate fault is a target map element labeling fault, and the target map element in the map information is detected to have the labeling fault, where the target map element includes at least one of a speed limit identifier, a lane line, a traffic indicator, and a lane;
the determining subunit 2022 is specifically further configured to determine that the second candidate fault is the target fault when the second candidate fault is a binding relation fault between map elements and the binding relation fault between map elements in the map information is detected, where the binding relation includes at least one of a binding relation between different lanes and a binding relation between traffic lights and lanes.
Optionally, the preset anomaly includes at least one of a collision, a braking deceleration exceeding a preset deceleration, a vehicle sway exceeding a preset amplitude, exiting an automatic driving mode, reversing, and running a red light.
It should be noted that, the fault detection device 200 for an automatic driving vehicle provided in this embodiment can implement all the technical solutions of the fault detection method embodiments for an automatic driving vehicle, so at least all the technical effects described above can be implemented, and the description thereof is omitted here.
In the technical scheme of the disclosure, the acquisition, storage, application and the like of the related user personal information all conform to the regulations of related laws and regulations, and the public sequence is not violated.
According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.
Fig. 4 illustrates a schematic block diagram of an example electronic device 400 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 4, the electronic device 400 includes a computing unit 401 that can perform various suitable actions and processes according to a computer program stored in a Read Only Memory (ROM) 402 or a computer program loaded from a storage unit 408 into a Random Access Memory (RAM) 403. In RAM 403, various programs and data required for the operation of device 400 may also be stored. The computing unit 401, ROM 402, and RAM 403 are connected to each other by a bus 404. An input/output (I/O) interface 405 is also connected to bus 404.
Various components in the electronic device 400 are connected to the I/O interface 405, including an input unit 406, such as a keyboard, a mouse, etc., an output unit 407, such as various types of displays, speakers, etc., a storage unit 408, such as a magnetic disk, optical disk, etc., and a communication unit 409, such as a network card, modem, wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The computing unit 401 may be a variety of general purpose and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 401 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the respective methods and processes described above, for example, a failure detection method of an autonomous vehicle. For example, in some embodiments, the fault detection method of an autonomous vehicle may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 400 via the ROM 402 and/or the communication unit 409. When the computer program is loaded into RAM 403 and executed by computing unit 401, one or more steps of the above-described failure detection method of the autonomous vehicle are performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the fault detection method of the autonomous vehicle by any other suitable means (e.g. by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be a special or general purpose programmable processor, operable to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user, for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback), and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a Local Area Network (LAN), a Wide Area Network (WAN), and the Internet.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.