Data processing method and device based on automatic driving, medium, equipment and vehicle

Abstract

The invention discloses a data processing method and device based on automatic driving, a medium, equipment and a vehicle, and the method comprises the steps: detecting whether an own vehicle meets a driving data screening triggering condition or not in the process that the own vehicle carries out automatic driving based on a pre-trained decision planning model and vehicle driving data is recorded in real time, the driving data screening triggering condition comprises a triggering condition for screening an abnormal driving behavior; when the vehicle meets the driving data screening triggering condition, a first vehicle driving data set is triggered and screened out, the first vehicle driving data set comprises multiple frames of vehicle driving data recorded in a preset time period, and the preset time period comprises a period of time before and after the vehicle meets the driving data screening triggering condition; simulation training is carried out on the first vehicle driving data set based on an automatic driving decision planning rule, a second vehicle driving data set is obtained, and the second vehicle driving data set meets a safe driving strategy; and training the decision planning model by using the second vehicle driving data set.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and more specifically, to a data processing method, device, medium, equipment and vehicle based on autonomous driving. Background Art

[0002] AI (Artificial Intelligence) is widely used in autonomous driving, and it is one of the key technologies to achieve autonomous driving. Autonomous driving vehicles need to collect surrounding information through sensors, such as radar point clouds collected by lidar and images collected by cameras, and then use perception and recognition algorithms to process this information, identify objects such as roads, vehicles, pedestrians, and traffic lights, and then use decision-making and planning models to make decisions and plan driving routes, such as choosing the right lane, maintaining a safe distance, and avoiding collisions. Among them, when training the decision-making and planning model, it is usually necessary to collect driving data of experts in various driving scenarios as a training set to train the decision-making and planning model. In this case, the higher the quality requirements of the decision-making and planning model, the higher the labor cost of collecting expert driving data. Summary of the invention

[0003] The present application provides a data processing method, device, medium, equipment and vehicle based on autonomous driving, which can reduce the cost of collecting vehicle driving data while improving the quality of the decision-making planning model.

[0004] The specific technical solutions are as follows:

[0005] In a first aspect, an embodiment of the present application provides a data processing method based on autonomous driving, the method comprising:

[0006] In the process of the self-vehicle performing automatic driving based on the pre-trained decision planning model and recording the vehicle driving data in real time, detecting whether the self-vehicle meets the driving data screening trigger condition, wherein the driving data screening trigger condition includes a trigger condition for screening abnormal driving behavior;

[0007] When the vehicle meets the driving data screening trigger condition, a first vehicle driving data set is triggered to be screened out, wherein the first vehicle driving data set includes multiple frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition;

[0008] Performing simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving to obtain the second vehicle driving data set, wherein the second vehicle driving data set satisfies a safe driving strategy;

[0009] The decision planning model is trained using the second vehicle driving data set.

[0010] Through the above scheme, it can be seen that in the embodiment of the present application, when the vehicle performs automatic driving based on the pre-trained decision planning model and records the vehicle driving data in real time, when it is detected that the vehicle meets the triggering condition for screening abnormal driving behavior, the first vehicle driving data set is triggered to be screened out, and the first vehicle driving data set is simulated and trained based on the decision planning rules of automatic driving to obtain a second vehicle driving data set that meets the safe driving strategy. It can be seen that after the embodiment of the present application obtains a version of the decision planning model based on expert driving data training in advance, the decision planning model can be applied to the actual vehicle to collect abnormal driving behavior data, and after the abnormal driving behavior data is corrected, the decision planning model can be improved without having to find an expert to drive manually, thereby reducing labor costs.

[0011] In a first possible implementation manner of the first aspect, detecting whether the vehicle satisfies a driving data screening trigger condition includes:

[0012] When a manual trigger instruction is received, determining that the vehicle satisfies the driving data screening trigger condition; and / or,

[0013] When the current driving state of the vehicle satisfies a preset driving data screening rule, determining that the vehicle satisfies the driving data screening trigger condition, wherein the preset driving data screening rule includes a rule determined based on dangerous driving behaviors in different driving scenarios; and / or,

[0014] During the driving of the self-vehicle based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

[0015] It can be seen from the above scheme that the embodiment of the present application can manually trigger the screening of vehicle driving data, can automatically trigger the screening of vehicle driving data based on rules, and can automatically trigger the screening of vehicle driving data based on a takeover operation, thereby collecting abnormal driving behavior data from multiple aspects, and obtaining driving data that complies with safe driving strategies by correcting abnormal driving behavior, thereby efficiently completing the collection of vehicle driving data for the driving scenarios of concern.

[0016] In a second possible implementation manner of the first aspect, the preset driving data screening rule includes at least one of the following:

[0017] The collision time TTC between the front of the vehicle and the rear of the target vehicle is less than or equal to the first time threshold;

[0018] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to a first distance threshold;

[0019] During the lane change process of the vehicle, the distance between the vehicle and surrounding vehicles is less than or equal to a second distance threshold;

[0020] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to a third distance threshold;

[0021] During the process of the ego vehicle avoiding the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to a second time threshold.

[0022] In a third possible implementation manner of the first aspect, before performing simulation training on the first vehicle driving dataset based on the decision planning rule for autonomous driving to obtain the second vehicle driving dataset, the method further includes:

[0023] According to a preset rule, metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set is determined, and the first vehicle driving data set and the metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set constitute an updated first vehicle driving data set, wherein the metadata information includes additional information for describing the vehicle driving data;

[0024] The step of performing simulation training on the first vehicle driving data set based on the autonomous driving decision planning rule to obtain the second vehicle driving data set includes:

[0025] The updated first vehicle driving data set is simulated and trained based on the decision planning rule to obtain the second vehicle driving data set.

[0026] It can be seen from the above scheme that the embodiment of the present application adds metadata information to assist model training on the basis of the original vehicle driving data, thereby further improving the accuracy of the decision-making planning model.

[0027] In a fourth possible implementation manner of the first aspect, the preset rule includes a rule determined based on the decision planning rule.

[0028] In a second aspect, an embodiment of the present application provides a data processing device based on autonomous driving, the device comprising:

[0029] A detection unit, used to detect whether the self-vehicle meets a driving data screening trigger condition when the self-vehicle is driving automatically based on a pre-trained decision planning model and recording vehicle driving data in real time, wherein the driving data screening trigger condition includes a trigger condition for screening abnormal driving behavior;

[0030] A screening unit, configured to trigger screening of a first vehicle driving data set when the vehicle meets a driving data screening trigger condition, wherein the first vehicle driving data set includes a plurality of frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition;

[0031] a simulation unit, configured to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving to obtain the second vehicle driving data set, wherein the second vehicle driving data set satisfies a safe driving strategy;

[0032] A training unit is used to train the decision planning model using the second vehicle driving data set.

[0033] In a first possible implementation manner of the second aspect, the detection unit is used to:

[0034] When a manual trigger instruction is received, determining that the vehicle satisfies the driving data screening trigger condition; and / or,

[0035] When the current driving state of the vehicle satisfies a preset driving data screening rule, determining that the vehicle satisfies the driving data screening trigger condition, wherein the preset driving data screening rule includes a rule determined based on dangerous driving behaviors in different driving scenarios; and / or,

[0036] During the driving of the self-vehicle based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

[0037] In a second possible implementation manner of the second aspect, the preset driving data screening rule includes at least one of the following:

[0038] The collision time TTC between the front of the vehicle and the rear of the target vehicle is less than or equal to the first time threshold;

[0039] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to a first distance threshold;

[0040] During the lane change process of the vehicle, the distance between the vehicle and surrounding vehicles is less than or equal to a second distance threshold;

[0041] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to a third distance threshold;

[0042] During the process of the ego vehicle avoiding the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to a second time threshold.

[0043] In a third possible implementation manner of the second aspect, the method further includes:

[0044] a determining unit, configured to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving, and before obtaining the second vehicle driving data set, determine metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set according to a preset rule, and form an updated first vehicle driving data set with the first vehicle driving data set and the metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set, wherein the metadata information includes additional information for describing the vehicle driving data;

[0045] The simulation unit is used to perform simulation training on the updated first vehicle driving data set based on the decision planning rule to obtain the second vehicle driving data set.

[0046] In a fourth possible implementation manner of the second aspect, the preset rule includes a rule determined based on the decision planning rule.

[0047] Through the above scheme, it can be seen that in the embodiment of the present application, when the vehicle performs automatic driving based on the pre-trained decision planning model and records the vehicle driving data in real time, when it is detected that the vehicle meets the triggering condition for screening abnormal driving behavior, the first vehicle driving data set is triggered to be screened out, and the first vehicle driving data set is simulated and trained based on the decision planning rules of automatic driving to obtain a second vehicle driving data set that meets the safe driving strategy. It can be seen that after the embodiment of the present application obtains a version of the decision planning model based on expert driving data training in advance, the decision planning model can be applied to the actual vehicle to collect abnormal driving behavior data, and after the abnormal driving behavior data is corrected, the decision planning model can be improved without having to find an expert to drive manually, thereby reducing labor costs.

[0048] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any possible implementation manner of the first aspect.

[0049] In a fourth aspect, an embodiment of the present application provides an electronic device, the electronic device comprising:

[0050] one or more processors;

[0051] The processor is coupled to a storage device, and the storage device is used to store one or more programs;

[0052] When one or more programs are executed by one or more processors, the electronic device implements the method described in any possible implementation manner of the first aspect.

[0053] In a fifth aspect, an embodiment of the present application provides a vehicle, which includes a device as described in any implementation of the second aspect, or includes an electronic device as described in the fourth aspect.

[0054] In a sixth aspect, an embodiment of the present application provides a computer program product, which includes instructions. When the instructions are executed on a computer or a processor, the computer or the processor executes the method described in any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art description are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work.

[0056] Figure 1 A flowchart of a data processing method based on autonomous driving provided in an embodiment of the present application;

[0057] Figure 2 A block diagram of a data processing device based on autonomous driving provided in an embodiment of the present application;

[0058] Figure 3 A schematic diagram of the structure of an electronic device or computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The terms "including" and "having" in the embodiments of the present application and the accompanying drawings and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.

[0061] Figure 1 The present invention is a flowchart of a data processing method based on autonomous driving, which can be applied to electronic devices or computer devices, and can be applied to terminals or servers. Specifically, the method may include the following steps:

[0062] S110: When the self-vehicle performs automatic driving based on a pre-trained decision-making planning model and records vehicle driving data in real time, detecting whether the self-vehicle satisfies a driving data screening trigger condition.

[0063] After a version of the decision planning model is trained using the collected expert driving data, the decision planning model can be applied to multiple vehicles, and each vehicle can be driven automatically using the decision planning model in different driving scenarios. During the process of each vehicle driving automatically based on the pre-trained decision planning model, the data acquisition device on the vehicle will collect relevant data in real time. For example, the camera will collect road images in front of / around the vehicle in real time, the positioning system will locate the vehicle in real time, and the lidar will collect lidar point clouds in front of / around the vehicle in real time. The collected data will also be processed to obtain raw data information as vehicle driving data.

[0064] Among them, the vehicle driving data is objective data obtained by processing the raw data information collected by the data acquisition device on the vehicle. The vehicle driving data includes the vehicle information, other vehicle information, static landmark information, traffic light information, etc. The vehicle information includes the real-time global position of the vehicle, the real-time global orientation angle, the real-time vehicle speed, the real-time following gear position, the target vehicle speed of the vehicle, etc. The other vehicle information includes the real-time global position, the real-time global orientation angle, the real-time vehicle speed, the real-time following gear position, etc. of other vehicles around the vehicle. The static landmark information includes the static landmark information on the road surface within a preset distance range around the vehicle, such as lane lines, road boundary lines, static obstacles, etc. Traffic light information includes the type and status of traffic lights. The types of traffic lights include motor vehicle traffic lights, non-motor vehicle traffic lights, lane lights, flashing warning lights, pedestrian crossing lights, direction indicators, road and railway level crossing lights, etc. The status of traffic lights includes the on and off status of traffic lights. The target speed of the vehicle is the expected speed set by the user.

[0065] During the process of the self-driving vehicle, it can be detected whether the self-vehicle meets the driving data screening trigger condition. When it is detected that the self-vehicle meets the driving data screening trigger condition, it can be triggered to screen the required vehicle driving data from the real-time recorded vehicle driving data. Among them, the driving data screening trigger condition includes the trigger condition for screening abnormal driving behavior.

[0066] The method for detecting whether the vehicle meets the driving data screening trigger condition includes at least one of the following:

[0067] The first method: when a manual trigger command is received, it is determined that the vehicle meets the driving data screening trigger condition.

[0068] During the autonomous driving process, when the vehicle is about to be in danger, or the driving behavior of the vehicle does not comply with expert-level safe driving rules, the driver can initiate a manual trigger command. For example, a trigger button can be set in advance in the autonomous driving operation interface of the vehicle (which can be the central control screen of the vehicle or the operation interface on a mobile terminal that interacts with the vehicle) for the driver to click to trigger.

[0069] The second method: when the current driving state of the vehicle satisfies the preset driving data screening rules, it is determined that the vehicle satisfies the driving data screening trigger condition.

[0070] The preset driving data screening rules include rules for dangerous driving behaviors based on different driving scenarios, including navigation lane change, suppressed lane change, detour lane change, overtaking lane change, intersection passage, speed control, lane avoidance, lane centering, and avoidance passage.

[0071] The preset driving data screening rules include at least one of the following:

[0072] The TTC (Time To Collision) between the front of the ego vehicle and the rear of the target vehicle is less than or equal to the first time threshold;

[0073] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to the first distance threshold;

[0074] During the lane change process of the ego vehicle, the distance between the ego vehicle and surrounding vehicles is less than or equal to the second distance threshold;

[0075] When the ego vehicle is avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to the third distance threshold;

[0076] When the ego vehicle avoids the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to the second time threshold.

[0077] The target vehicle may be a large vehicle having a size significantly different from that of the vehicle itself, such as a truck.

[0078] The values ​​of the first time threshold, the first distance threshold, the second distance threshold, the third distance threshold and the second time threshold can be determined according to actual experience. For example, the first time threshold can be 3s, the first distance threshold can be 0.5m, the second distance threshold can be 0.4m, the third distance threshold can be 0.3m, and the second time threshold can be 1s.

[0079] The third method: when the self-vehicle is driving based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

[0080] When the self-driving car is driving based on the automatic driving system, the driver will take over when danger is about to occur, and the self-driving car will switch from automatic driving to manual driving. Therefore, the vehicle driving data generated within a period of time (such as 15 seconds) before and after the takeover operation can be called vehicle driving data of abnormal driving behavior. Therefore, when the takeover operation is detected, it is determined that the self-driving car meets the driving data screening trigger condition.

[0081] S120: When the vehicle meets the driving data screening trigger condition, a first vehicle driving data set is screened out.

[0082] The first vehicle driving data set includes multiple frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition, such as 15 seconds before and after. When the vehicle does not meet the driving data screening trigger condition, there is no need to trigger the screening of the first vehicle driving data set, and the autonomous driving continues based on the pre-trained decision planning model.

[0083] S130: Perform simulation training on the first vehicle driving data set based on the decision planning rule of autonomous driving to obtain a second vehicle driving data set.

[0084] Among them, the second vehicle driving data set satisfies the safe driving strategy. The safe driving strategy is a driving strategy formulated based on safe driving behavior, and the driving behavior that does not meet the safe driving behavior is the abnormal driving behavior mentioned above. The embodiment of the present application may not limit the specific content of the safe driving strategy, as long as it can meet the safe driving requirements.

[0085] In the related art, decision planning can be performed by matching decision planning rules based on autonomous driving, and the setting of decision planning rules satisfies the safe driving strategy. In the embodiment of the present application, the safe driving strategy may include decision planning rules. After multiple rounds of simulation training of the first vehicle driving data set based on the decision planning rules of autonomous driving, the second vehicle driving data set can be made to meet the decision planning rules based on autonomous driving. That is to say, the embodiment of the present application can use simulation training to make the first vehicle driving data set that originally does not meet the decision planning rules based on autonomous driving become the second vehicle driving data set that meets the decision planning rules.

[0086] "Decision-making planning rules" refer to traditional planning algorithms based on optimization, which are rules pre-set one by one based on driving experience. During autonomous driving, decision-making planning can be performed in a manner that matches the decision-making planning rules. The pre-set decision-making planning rules themselves will not change as the actual driving scenarios are automatically learned. Therefore, the decision-making planning rules are different from the planning algorithms based on deep learning, that is, they are different from the decision-making planning models trained based on driving data in the embodiments of the present application.

[0087] The embodiment of the present application can perform an automatic driving simulation on the first vehicle driving data set based on the decision planning rule, and use the decision planning rule as a learning standard, and continuously adjust the abnormal driving data (such as driving trajectory, driving speed, etc.) in the first vehicle driving data set so that the adjusted abnormal driving data meets the decision planning rule, thereby achieving a correction effect on the abnormal driving data. In the embodiment of the present application, an automatic driving simulation is performed based on the decision planning rule and the first vehicle driving data set to obtain a driving trajectory that meets the driving requirements of the driving scene corresponding to the first vehicle driving data set; the driving trajectory that meets the driving requirements of the driving scene replaces the corresponding driving trajectory in the first vehicle driving data set that does not meet the driving requirements of the driving scene, and obtains a second vehicle driving data set.

[0088] S140: Training the decision planning model using the second vehicle driving data set.

[0089] After creating a corresponding second vehicle driving data set that satisfies the safe driving strategy based on the first vehicle driving data set including abnormal driving behavior, the pre-trained decision planning model can be trained based on the second vehicle driving data set to improve the decision planning accuracy and safety of the decision planning model.

[0090] The data processing method based on automatic driving provided in the embodiment of the present application, when the vehicle performs automatic driving based on the pre-trained decision planning model and records the vehicle driving data in real time, when it is detected that the vehicle meets the triggering condition for screening abnormal driving behavior, the first vehicle driving data set is triggered to be screened out, and the first vehicle driving data set is simulated and trained based on the decision planning rules of automatic driving to obtain a second vehicle driving data set that meets the safe driving strategy. It can be seen that after the embodiment of the present application obtains a version of the decision planning model based on expert driving data training in advance, the decision planning model can be applied to the actual vehicle to collect abnormal driving behavior data, and after the abnormal driving behavior data is corrected, the decision planning model can be improved without having to find an expert to drive manually, thereby reducing labor costs.

[0091] In one embodiment, before the first vehicle driving data set is simulated and trained based on the decision planning rules of autonomous driving to obtain the second vehicle driving data set, metadata information corresponding to each frame of vehicle driving data in the first vehicle driving data set is determined according to preset rules, and the first vehicle driving data set and the metadata information corresponding to each frame of vehicle driving data in the first vehicle driving data set constitute an updated first vehicle driving data set, and then the updated first vehicle driving data set is simulated and trained based on the decision planning rules to obtain the second vehicle driving data set.

[0092] The metadata information includes additional information for describing vehicle driving data. The metadata information includes the curvature of the current road, diverging flow information, the lateral distance of the vehicle from the current lane, the intention of the vehicle to change lanes, relevant information of the vehicle in front of the left lane, relevant information of the vehicle in front of the right lane, relevant information of the vehicle behind the left lane, relevant information of the vehicle behind the right lane, whether there is a vehicle in front of the current lane, relevant information of the vehicle in front of the current lane when there is a vehicle in front of the current lane, road speed limit information, TTC of the vehicle and other vehicles during the vehicle's lane change, and TTC of other vehicles and the vehicle during the vehicle's lane change. The metadata information corresponding to each frame of vehicle driving data may include one type of information or multiple types of information, that is, one frame of vehicle driving data corresponds to at least one type of metadata information.

[0093] The method for determining metadata information corresponding to each frame of vehicle driving data in a first vehicle driving data set according to preset rules includes: determining the preset rules according to decision-making planning rules based on automatic driving; and determining metadata information corresponding to each frame of vehicle driving data in the first vehicle driving data set according to the preset rules.

[0094] In the embodiment of the present application, the user can customize the preset rules, or the preset rules can be automatically determined by using the existing decision planning rules. The embodiment of the present application can determine the preset rules based on the decision planning rules, such as extracting the key information in the decision planning rules, and using the calculation method of the key information as a preset rule. The key information includes some customized data in addition to the objective data in the vehicle driving data. For example, the vehicle driving data only contains the information of the vehicle itself and other vehicles, but no information related to the vehicle itself and other vehicles, such as the concepts of the front vehicle in the left lane and the rear vehicle in the left lane. Therefore, when this type of information exists in the decision planning rules, the key information can be extracted from the decision planning rules, and the calculation method of the key information can be used as a preset rule. The calculation method of the key information can refer to the prior art and will not be repeated here.

[0095] It can be seen from the above scheme that the embodiment of the present application adds metadata information to assist model training on the basis of the original vehicle driving data, thereby further improving the accuracy of the decision-making planning model.

[0096] Corresponding to the above method embodiment, another embodiment of the present application provides a data processing device based on autonomous driving, such as Figure 2 As shown, the device comprises:

[0097] A detection unit 210 is used to detect whether the self-vehicle meets a driving data screening trigger condition when the self-vehicle is driving automatically based on a pre-trained decision planning model and recording vehicle driving data in real time, wherein the driving data screening trigger condition includes a trigger condition for screening abnormal driving behavior;

[0098] A screening unit 220, configured to trigger screening of a first vehicle driving data set when the vehicle meets a driving data screening trigger condition, wherein the first vehicle driving data set includes a plurality of frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition;

[0099] A simulation unit 230 is used to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving to obtain the second vehicle driving data set, wherein the second vehicle driving data set satisfies a safe driving strategy;

[0100] The training unit 240 is used to train the decision planning model using the second vehicle driving data set.

[0101] In a possible implementation, the detection unit 210 is configured to:

[0102] When a manual trigger instruction is received, determining that the vehicle satisfies the driving data screening trigger condition; and / or,

[0103] When the current driving state of the vehicle satisfies a preset driving data screening rule, determining that the vehicle satisfies the driving data screening trigger condition, wherein the preset driving data screening rule includes a rule determined based on dangerous driving behaviors in different driving scenarios; and / or,

[0104] During the driving of the self-vehicle based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

[0105] In a possible implementation, the preset driving data screening rule includes at least one of the following:

[0106] The collision time TTC between the front of the vehicle and the rear of the target vehicle is less than or equal to the first time threshold;

[0107] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to a first distance threshold;

[0108] During the lane change process of the vehicle, the distance between the vehicle and surrounding vehicles is less than or equal to a second distance threshold;

[0109] During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to a third distance threshold;

[0110] During the process of the ego vehicle avoiding the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to a second time threshold.

[0111] In a possible implementation, the method further includes:

[0112] a determining unit, configured to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving, and before obtaining the second vehicle driving data set, determine metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set according to a preset rule, and form an updated first vehicle driving data set with the first vehicle driving data set and the metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set, wherein the metadata information includes additional information for describing the vehicle driving data;

[0113] The simulation unit 230 is used to perform simulation training on the updated first vehicle driving data set based on the decision planning rule to obtain the second vehicle driving data set.

[0114] In a possible implementation manner, the preset rule includes a rule determined based on the decision planning rule.

[0115] The data processing device based on automatic driving provided by the embodiment of the present application, when the vehicle performs automatic driving based on the pre-trained decision planning model and records the vehicle driving data in real time, triggers the screening of the first vehicle driving data set when it is detected that the vehicle meets the triggering condition for screening abnormal driving behavior, and simulates the first vehicle driving data set based on the decision planning rules of automatic driving to obtain the second vehicle driving data set that meets the safe driving strategy. It can be seen that after the embodiment of the present application obtains a version of the decision planning model based on expert driving data training in advance, the decision planning model can be applied to the actual vehicle to collect abnormal driving behavior data, and after the abnormal driving behavior data is corrected, the decision planning model can be improved without having to find an expert to drive manually, thereby reducing labor costs.

[0116] Based on the above method embodiments, another embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any of the above embodiments is implemented.

[0117] Based on the above method embodiment, another embodiment of the present application provides an electronic device or a computer device, such as Figure 3 As shown, including:

[0118] One or more processors 310;

[0119] The processor 310 is coupled to a storage device 320, and the storage device 320 is used to store one or more programs;

[0120] When the one or more programs are executed by the one or more processors 310, the electronic device or computer device implements the method described in any of the above embodiments.

[0121] Based on the above method embodiments, another embodiment of the present application provides a vehicle, which includes the device as described in any of the above embodiments, or includes the electronic device as described above.

[0122] The vehicle includes a CPU (Central Processing Unit), a T-Box (Telematics Box), an image collector, a positioning device, and a laser radar. Among them, the image collector is used to collect road environment images, the positioning device is used to locate the vehicle, the laser radar is used to collect radar point clouds, and the CPU is used to process road environment images, vehicle positioning information, and radar point clouds, obtain and record vehicle driving data, and execute any of the above-mentioned embodiments of the data processing method based on automatic driving. The CPU can also send the first vehicle driving data set that is triggered and filtered out to the server through the T-Box, and the server simulates and trains the first vehicle driving data set based on the decision-making planning rules of automatic driving to obtain a second vehicle driving data set, and uses the second vehicle driving data set to train the decision-making planning model.

[0123] Based on the above embodiments, another embodiment of the present application provides a computer program product, which includes instructions. When the instructions are executed on a computer or a processor, the computer or the processor executes the method described in any of the above embodiments.

[0124] The above device embodiment corresponds to the method embodiment and has the same technical effect as the method embodiment. For specific description, please refer to the method embodiment. The device embodiment is obtained based on the method embodiment. For specific description, please refer to the method embodiment part, which will not be repeated here. It can be understood by those skilled in the art that the accompanying drawings are only schematic diagrams of one embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present application.

[0125] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A data processing method based on autonomous driving, characterized in that: The method comprises: In the process of the self-vehicle performing automatic driving based on the pre-trained decision planning model and recording the vehicle driving data in real time, detecting whether the self-vehicle meets the driving data screening trigger condition, wherein the driving data screening trigger condition includes a trigger condition for screening abnormal driving behavior; When the vehicle meets the driving data screening trigger condition, a first vehicle driving data set is triggered to be screened out, wherein the first vehicle driving data set includes multiple frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition; Performing simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving to obtain the second vehicle driving data set, wherein the second vehicle driving data set satisfies a safe driving strategy; The decision planning model is trained using the second vehicle driving data set.

2. The method according to claim 1, characterized in that The detecting whether the vehicle satisfies the driving data screening trigger condition includes: When a manual trigger instruction is received, determining that the vehicle satisfies the driving data screening trigger condition; and / or, When the current driving state of the vehicle satisfies a preset driving data screening rule, determining that the vehicle satisfies the driving data screening trigger condition, wherein the preset driving data screening rule includes a rule determined based on dangerous driving behaviors in different driving scenarios; and / or, During the driving of the self-vehicle based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

3. The method according to claim 2, characterized in that The preset driving data screening rule includes at least one of the following: The collision time TTC between the front of the vehicle and the rear of the target vehicle is less than or equal to the first time threshold; During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to a first distance threshold; During the lane change process of the vehicle, the distance between the vehicle and surrounding vehicles is less than or equal to a second distance threshold; During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to a third distance threshold; During the process of the ego vehicle avoiding the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to a second time threshold.

4. The method according to claim 1, characterized in that: Before performing simulation training on the first vehicle driving data set based on the decision planning rule of autonomous driving to obtain the second vehicle driving data set, the method further includes: According to a preset rule, metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set is determined, and the first vehicle driving data set and the metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set constitute an updated first vehicle driving data set, wherein the metadata information includes additional information for describing the vehicle driving data; The step of performing simulation training on the first vehicle driving data set based on the autonomous driving decision planning rule to obtain the second vehicle driving data set includes: The updated first vehicle driving data set is simulated and trained based on the decision planning rule to obtain the second vehicle driving data set.

5. The method according to any one of claims 1 to 4, characterized in that The preset rules include rules determined based on the decision planning rules.

6. A data processing device based on automatic driving, characterized in that: The device comprises: A detection unit, used to detect whether the self-vehicle meets a driving data screening trigger condition when the self-vehicle is driving automatically based on a pre-trained decision planning model and recording vehicle driving data in real time, wherein the driving data screening trigger condition includes a trigger condition for screening abnormal driving behavior; A screening unit, configured to trigger screening of a first vehicle driving data set when the vehicle meets a driving data screening trigger condition, wherein the first vehicle driving data set includes a plurality of frames of vehicle driving data recorded within a preset time period, and the preset time period includes a period of time before and after the vehicle meets the driving data screening trigger condition; a simulation unit, configured to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving to obtain the second vehicle driving data set, wherein the second vehicle driving data set satisfies a safe driving strategy; A training unit is used to train the decision planning model using the second vehicle driving data set.

7. The device according to claim 6, characterized in that The detection unit is used for: When a manual trigger instruction is received, determining that the vehicle satisfies the driving data screening trigger condition; and / or, When the current driving state of the vehicle satisfies a preset driving data screening rule, determining that the vehicle satisfies the driving data screening trigger condition, wherein the preset driving data screening rule includes a rule determined based on dangerous driving behaviors in different driving scenarios; and / or, During the driving of the self-vehicle based on the automatic driving system, when a takeover operation is detected, it is determined that the self-vehicle meets the driving data screening trigger condition.

8. The device according to claim 7, characterized in that The preset driving data screening rule includes at least one of the following: The collision time TTC between the front of the vehicle and the rear of the target vehicle is less than or equal to the first time threshold; During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the target vehicle is less than or equal to a first distance threshold; During the lane change process of the vehicle, the distance between the vehicle and surrounding vehicles is less than or equal to a second distance threshold; During the process of the ego vehicle avoiding the target vehicle, the distance between the ego vehicle and the lane line is less than or equal to a third distance threshold; During the process of the ego vehicle avoiding the target vehicle, the shortest avoidance time after the ego vehicle passes the target vehicle is less than or equal to a second time threshold.

9. The device according to claim 6, characterized in that The method further comprises: a determining unit, configured to perform simulation training on the first vehicle driving data set based on a decision-making planning rule for autonomous driving, and before obtaining the second vehicle driving data set, determine metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set according to a preset rule, and form an updated first vehicle driving data set with the first vehicle driving data set and the metadata information corresponding to each frame of the vehicle driving data in the first vehicle driving data set, wherein the metadata information includes additional information for describing the vehicle driving data; The simulation unit is used to perform simulation training on the updated first vehicle driving data set based on the decision planning rule to obtain the second vehicle driving data set.

10. The device according to any one of claims 6 to 9, characterized in that: The preset rules include rules determined based on the decision planning rules.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

12. An electronic device, characterized in that: The electronic device comprises: one or more processors; The processor is coupled to a storage device, and the storage device is used to store one or more programs; When the one or more programs are executed by the one or more processors, the electronic device implements the method according to any one of claims 1 to 5.

13. A vehicle, characterized in that: The vehicle comprises the device according to any one of claims 6 to 10, or comprises the electronic device according to claim 12.

Images