Congested lane-changing simulation scenario evaluation method and related equipment

Abstract

The present invention relates to the field of autonomous driving technology and discloses a congested lane change simulation scenario evaluation method and related equipment for improving the accuracy and efficiency of test scoring for autonomous driving software handling congested lane change scenarios, as well as improving the safety of autonomous driving. The congested lane change simulation scenario evaluation method comprises: obtaining actual vehicle driving scene information corresponding to a target vehicle, segmenting the actual vehicle driving scene information into congested lane change scenarios, and obtaining multiple congested lane change actual scene information; performing an autonomous driving behavior test based on the multiple congested lane change actual scene information and turn signal flashing information, and obtaining autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario; and performing lane change behavior similarity scoring processing based on the multiple congested lane change actual scene information and the autonomous driving vehicle lane change simulation test information, and obtaining a simulation scenario test score result, wherein the lane change behavior includes the lane change trajectory and the lane change start time.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method for evaluating a crowded lane change simulation scenario and related equipment. Background Art

[0002] Lane changes, also known as lane changes, occur when a car moves from one lane to another in the same direction, such as to overtake or avoid an obstacle. Before changing lanes, a lane change signal must be issued to alert other vehicles. Congested lane changes are one of the most challenging scenarios for autonomous driving software.

[0003] In the existing technology, the evaluation method for autonomous driving software's handling of crowded lane change scenarios is mainly achieved through actual road testing. However, the road testing cycle is long, and because different versions of autonomous driving software encounter different crowded lane change scenarios in actual road testing, the scenario test scoring accuracy is low and the efficiency is low. Summary of the Invention

[0004] The present invention provides a crowded lane-changing simulation scenario evaluation method and related equipment, which are used to improve the test scoring accuracy and efficiency of autonomous driving software in handling crowded lane-changing scenarios, and to improve the safety of autonomous driving.

[0005] To achieve the above-mentioned objectives, the first aspect of the present invention provides a method for evaluating a crowded lane change simulation scenario, comprising: obtaining actual vehicle driving scene information corresponding to a target vehicle, the actual vehicle driving scene information including position information, speed information, turn signal flashing information, and obstacle information around the target vehicle passed by the target vehicle; performing crowded lane change scene segmentation on the actual vehicle driving scene information to obtain multiple crowded lane change actual scene information; performing an autonomous driving behavior test based on the multiple crowded lane change actual scene information and the turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario; performing lane change behavior similarity scoring processing based on the multiple crowded lane change actual scene information and the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario to obtain a simulation scenario test scoring result, the lane change behavior including the lane change trajectory and the lane change start time.

[0006] In a feasible implementation manner, the actual vehicle driving scene information is subjected to congested lane change scene segmentation to obtain multiple congested lane change actual scene information, including: performing scene segmentation processing on the actual vehicle driving scene information through a preset scene segmentation model to obtain a scene segmentation result; and scene filtering the scene segmentation result according to a preset congested lane change scene screening rule to obtain multiple congested lane change actual scene information.

[0007] In a feasible implementation, the autonomous driving behavior test is performed based on the multiple congested lane change actual scene information and the turn signal flashing information to obtain the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene, including: running the autonomous driving simulation algorithm to be tested respectively according to the multiple congested lane change actual scene information to obtain multiple congested lane change simulation scenes; reading the congested lane change moment and the corresponding congested lane change direction corresponding to each congested lane change simulation scene from the turn signal flashing information; in each congested lane change simulation scene, sending a lane change request toward the congested lane change direction to the autonomous driving vehicle according to the congested lane change moment corresponding to each congested lane change simulation scene to obtain the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene.

[0008] In a feasible implementation, the lane change behavior similarity scoring process is performed based on the multiple congested lane change actual scene information and the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene to obtain the simulation scene test scoring result, including: obtaining the target vehicle lane change information corresponding to each congested lane change simulation scene from the multiple congested lane change actual scene information; determining the target vehicle lane change start time and target vehicle projection distance corresponding to each congested lane change simulation scene based on the target vehicle lane change information corresponding to each congested lane change simulation scene, the target vehicle projection distance being used to indicate when the congested lane change is completed, the target vehicle lane change start time and target vehicle projection distance corresponding to each congested lane change simulation scene The vehicle body length projected onto the target obstacle vehicle; according to the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario, determine the autonomous driving vehicle lane change start time and autonomous driving vehicle projection distance corresponding to each congested lane change simulation scenario, the autonomous driving vehicle projection distance is used to indicate the vehicle body length of the autonomous driving vehicle projected onto the target obstacle vehicle in a direction perpendicular to the lane line when the congested lane change is completed; perform lane change behavior similarity scoring processing based on the target vehicle lane change start time, the target vehicle projection distance, the autonomous driving vehicle lane change start time and the autonomous driving vehicle projection distance to obtain the simulation scenario test scoring result.

[0009] In one feasible implementation, the lane change behavior similarity scoring process based on the target vehicle lane change start time, the target vehicle projected distance, the autonomous vehicle lane change start time, and the autonomous vehicle projected distance to obtain a simulation scenario test scoring result includes: calculating a projection distance difference corresponding to each congested lane change simulation scenario based on the target vehicle projected distance and the autonomous vehicle projected distance, and obtaining a lane change completion score corresponding to each congested lane change simulation scenario based on the projection distance difference corresponding to each congested lane change simulation scenario; calculating a lane change time difference corresponding to each congested lane change simulation scenario based on the target vehicle lane change start time and the autonomous vehicle lane change start time, and obtaining a lane change time score corresponding to each congested lane change simulation scenario based on the lane change time difference corresponding to each congested lane change simulation scenario; setting a congested lane change scenario score corresponding to each congested lane change simulation scenario based on the lane change completion score and the lane change time score corresponding to each congested lane change simulation scenario; and summing the congested lane change scenario scores corresponding to each congested lane change simulation scenario to obtain a simulation scenario test scoring result.

[0010] In a feasible implementation manner, before obtaining the actual vehicle driving scene information corresponding to the target vehicle, the actual vehicle driving scene information including the position information, speed information, turn signal flashing information and obstacle information around the target vehicle passed by the target vehicle, the congested lane change simulation scene evaluation method also includes: obtaining initial vehicle driving scene information in an actual road driving scene, the initial vehicle driving scene information including vehicle perception information and vehicle video image information; performing information verification processing on the initial vehicle driving scene information to obtain a verification result; when the verification result is verification passed, performing data analysis, annotation and data packaging processing on the initial vehicle driving scene information to obtain the actual vehicle driving scene information corresponding to the target vehicle.

[0011] In a feasible implementation, after performing lane change behavior similarity scoring processing based on the multiple congested lane change actual scene information and the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene to obtain a simulation scene test scoring result, and the lane change behavior includes a lane change trajectory and a lane change start time, the congested lane change simulation scene evaluation method further includes: determining whether the simulation scene test scoring result meets a preset scoring threshold; if the simulation scene test scoring result meets the preset scoring threshold, setting a test pass label for the autonomous driving simulation algorithm to be tested, and performing version release processing on the autonomous driving simulation algorithm to be tested.

[0012] A second aspect of the present invention provides a crowded lane change simulation scenario evaluation device, comprising: a first acquisition module, used to obtain actual vehicle driving scene information corresponding to a target vehicle, the actual vehicle driving scene information including position information, speed information, turn signal flashing information and obstacle information around the target vehicle passed by the target vehicle; a segmentation module, used to perform crowded lane change scene segmentation on the actual vehicle driving scene information to obtain multiple crowded lane change actual scene information; a testing module, used to perform an autonomous driving behavior test based on the multiple crowded lane change actual scene information and the turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario; a scoring module, used to perform lane change behavior similarity scoring processing based on the multiple crowded lane change actual scene information and the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario to obtain a simulation scenario test scoring result, the lane change behavior including the lane change trajectory and the lane change start time.

[0013] In a feasible implementation, the segmentation module is specifically used to: perform scene segmentation processing on the actual vehicle driving scene information through a preset scene segmentation model to obtain a scene segmentation result; and perform scene filtering on the scene segmentation result according to a preset congested lane change scene screening rule to obtain multiple congested lane change actual scene information.

[0014] In a feasible implementation, the test module is specifically used to: run the autonomous driving simulation algorithm to be tested according to the multiple actual crowded lane change scene information respectively to obtain multiple crowded lane change simulation scenes; read the crowded lane change time and the corresponding crowded lane change direction corresponding to each crowded lane change simulation scene from the turn signal flashing information; in each crowded lane change simulation scene, send a lane change request to the autonomous driving vehicle in the crowded lane change direction according to the crowded lane change time corresponding to each crowded lane change simulation scene to obtain the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scene.

[0015] In one feasible embodiment, the scoring module further includes: an acquisition unit for acquiring target vehicle lane change information corresponding to each congested lane change simulation scenario from the multiple congested lane change actual scenario information; a first determination unit for determining, based on the target vehicle lane change information corresponding to each congested lane change simulation scenario, the target vehicle lane change start time and target vehicle projection distance corresponding to each congested lane change simulation scenario, the target vehicle projection distance being used to indicate the vehicle body length of the target vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed; a second determination unit for determining, based on the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario, the autonomous driving vehicle lane change start time and autonomous driving vehicle projection distance corresponding to each congested lane change simulation scenario, the autonomous driving vehicle projection distance being used to indicate the vehicle body length of the autonomous driving vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed; and a scoring unit for performing lane change behavior similarity scoring processing based on the target vehicle lane change start time, the target vehicle projection distance, the autonomous driving vehicle lane change start time, and the autonomous driving vehicle projection distance to obtain a simulation scenario test scoring result.

[0016] In a feasible implementation manner, the scoring unit is specifically used to: calculate the projection distance difference corresponding to each crowded lane change simulation scenario according to the target vehicle projection distance and the autonomous driving vehicle projection distance, and obtain the lane change completion score corresponding to each crowded lane change simulation scenario according to the projection distance difference corresponding to each crowded lane change simulation scenario; calculate the lane change time difference corresponding to each crowded lane change simulation scenario according to the target vehicle lane change start time and the autonomous driving vehicle lane change start time, and obtain the lane change time score corresponding to each crowded lane change simulation scenario according to the lane change time difference corresponding to each crowded lane change simulation scenario; set the crowded lane change scene score corresponding to each crowded lane change simulation scenario based on the lane change completion score corresponding to each crowded lane change simulation scenario and the lane change time score corresponding to each crowded lane change simulation scenario; summarize the crowded lane change scene scores corresponding to each crowded lane change simulation scenario to obtain a simulation scenario test scoring result.

[0017] In a feasible implementation manner, the crowded lane change simulation scenario evaluation device also includes: a second acquisition module, used to obtain initial vehicle driving scene information in an actual road driving scenario, wherein the initial vehicle driving scene information includes vehicle perception information and vehicle video image information; a verification module, used to perform information verification processing on the initial vehicle driving scene information to obtain a verification result; and a labeling module, used to perform data analysis, labeling and data packaging processing on the initial vehicle driving scene information when the verification result is verification passed, to obtain actual vehicle driving scene information corresponding to the target vehicle.

[0018] In a feasible implementation, the crowded lane change simulation scenario evaluation device further includes: a judgment module, configured to judge whether the simulation scenario test score result meets a preset score threshold; a setting module, configured to set a test pass label for the autonomous driving simulation algorithm to be tested if the simulation scenario test score result meets the preset score threshold, and perform version release processing on the autonomous driving simulation algorithm to be tested.

[0019] A third aspect of the present invention provides a crowded lane change simulation scenario evaluation device, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the crowded lane change simulation scenario evaluation device executes the above-mentioned crowded lane change simulation scenario evaluation method.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored therein, which, when executed on a computer, enables the computer to execute the above-mentioned method for evaluating a crowded lane change simulation scenario.

[0021] In the technical solution provided by the present invention, actual vehicle driving scene information corresponding to the target vehicle is obtained, and the actual vehicle driving scene information includes the position information, speed information, turn signal flashing information and obstacle information around the target vehicle passed by the target vehicle; the actual vehicle driving scene information is segmented into crowded lane change scenes to obtain multiple crowded lane change actual scene information; an automatic driving behavior test is performed based on the multiple crowded lane change actual scene information and the turn signal flashing information to obtain automatic driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scene; a lane change behavior similarity scoring process is performed based on the multiple crowded lane change actual scene information and the automatic driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scene to obtain a simulation scene test scoring result, and the lane change behavior includes a lane change trajectory and a lane change start time. In an embodiment of the present invention, an autonomous driving behavior test is performed using information from multiple actual crowded lane change scenarios and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. A lane change behavior similarity scoring process is performed based on the information from multiple actual crowded lane change scenarios and the information from multiple autonomous driving vehicle lane change simulation test information to obtain a simulation scenario test scoring result, thereby improving the test scoring accuracy and efficiency of the autonomous driving software in handling crowded lane change scenarios and improving the safety of autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of an embodiment of a method for evaluating a crowded lane-changing simulation scenario according to an embodiment of the present invention;

[0023] Figure 2A schematic diagram of a vehicle projection during a crowded lane change process according to an embodiment of the present invention;

[0024] Figure 3 A schematic diagram of another embodiment of a method for evaluating a crowded lane-changing simulation scenario according to an embodiment of the present invention;

[0025] Figure 4 Schematic diagram of a function between the difference in projected distance along the perpendicular lane line between the autonomous driving vehicle and the target vehicle and the lane change completion score in an embodiment of the present invention;

[0026] Figure 5 Schematic diagram of a function between the lane change start time difference between the autonomous driving vehicle and the target vehicle and the lane change time score in an embodiment of the present invention;

[0027] Figure 6 A schematic diagram of an embodiment of a device for evaluating a crowded lane-changing simulation scenario according to an embodiment of the present invention;

[0028] Figure 7 Schematic diagram of another embodiment of a device for evaluating a crowded lane-changing simulation scenario according to an embodiment of the present invention;

[0029] Figure 8 Schematic diagram of an embodiment of a crowded lane change simulation scenario evaluation device in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] Embodiments of the present invention provide a method and related equipment for evaluating a crowded lane-changing simulation scenario, which are used to improve the test scoring accuracy and efficiency of autonomous driving software in handling crowded lane-changing scenarios, and to improve the safety of autonomous driving.

[0031] The terms "first," "second," "third," "fourth," and the like (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatus.

[0032] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 An embodiment of a method for evaluating a crowded lane change simulation scenario in an embodiment of the present invention includes:

[0033] 101. Obtain actual vehicle driving scene information corresponding to the target vehicle. The actual vehicle driving scene information includes position information, speed information, turn signal flashing information, and obstacle information around the target vehicle.

[0034] The actual vehicle driving scene information includes the target vehicle's passing location information, speed information, turn signal flashing information, and information about obstacles around the target vehicle. The turn signal flashing information includes the turn signal activation time and turn signal direction. The turn signal direction includes left turn signal direction and right turn signal direction. The turn signal direction is used to indicate the target vehicle's lane change direction. It should be noted that when the target vehicle changes lanes in a congested scene, the positioning module, perception module, and data acquisition module corresponding to the target vehicle's autonomous driving system are simultaneously activated to record the target vehicle's passing location information, speed information, turn signal flashing information, and information about obstacles around the target vehicle in real time. Specifically, the server receives a congested lane change simulation scenario evaluation request and obtains a congested lane change scenario identifier from the congested lane change simulation scenario evaluation request. The server verifies the congested lane change scenario identifier according to preset scenario identification rules to obtain a verification result. When the verification result is successful, the server converts the congested lane change scenario identifier into a scenario query statement according to the syntax rules of the structured query language. The server executes the scenario query statement to retrieve the actual vehicle driving scene information corresponding to the target vehicle from the preset scenario database.

[0035] Before step 101, in some embodiments, the server obtains initial vehicle driving scene information in an actual road driving scenario, where the initial vehicle driving scene information includes vehicle perception information and vehicle video image information; the server performs information verification processing on the initial vehicle driving scene information to obtain a verification result; when the verification result is that the verification is passed, the server performs data analysis, annotation and data encapsulation processing on the initial vehicle driving scene information to obtain actual vehicle driving scene information corresponding to the target vehicle.

[0036] Furthermore, the server generates a congested lane-changing scene identifier, and stores actual vehicle driving scene information corresponding to the target vehicle into a preset scene database according to the congested lane-changing scene identifier.

[0037] It should be noted that the server performs data analysis and annotation on the initial vehicle driving scene information, including scene category annotation processing and scene segmentation time period annotation processing on the initial vehicle driving scene information, so that the server can execute step 102.

[0038] It is understandable that the execution subject of the present invention may be a crowded lane change simulation scenario assessment device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0039] 102. Perform congested lane change scene segmentation on the actual vehicle driving scene information to obtain multiple congested lane change actual scene information.

[0040] Among them, the actual scene information of each crowded lane change includes the scene of lane change between the target vehicle and the target obstacle vehicle. Before the lane change, the target vehicle and the target obstacle vehicle are traveling in parallel and in the same direction along two or more lanes. The lane change process is the process of the target vehicle occupying the lane in front of the target obstacle vehicle. After the lane change, the target vehicle and the target obstacle vehicle are traveling in the same lane.

[0041] In some embodiments, the server performs scene segmentation processing on the actual vehicle driving scene information through a preset scene segmentation model to obtain a scene segmentation result, wherein the actual vehicle driving scene information is scene information that has been pre-labeled through data analysis, and the actual vehicle driving scene information has multiple labeled scene categories and multiple labeled scene segmentation time periods. Specifically, the server performs scene segmentation processing on the actual vehicle driving scene information according to the multiple labeled scene segmentation time periods through the preset scene segmentation model to obtain a scene segmentation result; the server performs scene filtering on the scene segmentation result according to a preset congested lane changing scene screening rule to obtain multiple congested lane changing actual scene information. Specifically, the server filters multiple lane changing driving scenes from the scene segmentation result according to the preset congested lane changing scene screening rule, and obtains traffic flow status information corresponding to each lane changing driving scene. The server determines the congestion level corresponding to each lane changing driving scene according to the traffic flow status information corresponding to each lane changing driving scene, and determines multiple congested lane changing actual scene information from the multiple lane changing driving scenes according to the congestion level corresponding to each lane changing driving scene.

[0042] It should be noted that the scene segmentation results include merging driving scenes, lane changing driving scenes, and overtaking driving scenes. The preset congested lane changing scene screening rules are used to instruct the extraction of lane changing driving scenes from the scene segmentation results, and to perform secondary scene screening based on the traffic flow information in the lane changing driving scenes to obtain multiple actual congested lane changing scene information.

[0043] 103. Conduct an autonomous driving behavior test based on multiple actual crowded lane change scenario information and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario.

[0044] The autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario is used to indicate specific information during the lane change process of the autonomous driving vehicle. Specifically, the server runs the autonomous driving simulation algorithm to be tested on each actual congested lane change scenario information to obtain the congested lane change simulation scenario corresponding to each actual congested lane change scenario information; the server reads the congested lane change time corresponding to each congested lane change simulation scenario and the congested lane change direction corresponding to each congested lane change simulation scenario from the turn signal flashing information; in each congested lane change simulation scenario, the server sends a lane change request in the congested lane change direction to the autonomous driving vehicle according to the congested lane change time corresponding to each congested lane change simulation scenario, thereby controlling the autonomous driving vehicle to execute the lane change operation in accordance with the lane change request in the congested lane change direction, and obtains the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario output by the autonomous driving simulation algorithm to be tested.

[0045] 104. Perform lane change behavior similarity scoring based on multiple actual crowded lane change scenario information and autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario to obtain simulation scenario test scoring results. The lane change behavior includes the lane change trajectory and the lane change start time.

[0046] Specifically, the server analyzes and compares the lane-changing behavior of the target vehicle in each actual crowded lane-change scenario with the lane-changing behavior of the autonomous vehicle in each simulated crowded lane-change scenario to generate simulation scenario test scores. Lane-change behavior includes lane-change trajectory and lane-change start time. Each target vehicle and autonomous vehicle has a corresponding target obstacle vehicle. The target obstacle vehicle indicates the rear obstacle vehicle closest to the target vehicle and autonomous vehicle in their lanes when the target vehicle and autonomous vehicle complete their lane change.

[0047] like Figure 2 As shown, in each actual crowded lane change scenario, vehicle a is the target vehicle, and vehicle b is the target obstacle vehicle. In each simulated crowded lane change scenario, vehicle a is the autonomous vehicle, and vehicle b is the target obstacle vehicle. During the lane change, the projection of the target vehicle or autonomous vehicle on the front of the target obstacle vehicle is a line segment AB. The server records the length of AB corresponding to the target vehicle in each frame as L1, and the length of AB corresponding to the autonomous vehicle in each frame as L2.

[0048] Specifically, when the server detects that both the target vehicle and the autonomous driving vehicle have completed the lane change behavior, the server obtains the corresponding target vehicle projection distance (that is, the maximum value max_L1 of L1) from each of the multiple crowded lane change actual scene information, and obtains the corresponding autonomous driving vehicle projection distance (that is, the maximum value max_L2 of L2) from the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scene. The server obtains the target vehicle's lane change start time (that is, the moment when L1 starts to be greater than 0, recorded as t1) from each of the multiple crowded lane change actual scene information, and obtains the corresponding autonomous driving vehicle lane change time (that is, the moment when L2 starts to be greater than 0, recorded as t2) from the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scene. The server compares the target vehicle projection distance and the autonomous driving vehicle projection distance, as well as the target vehicle's lane change start time and the autonomous driving vehicle lane change time. Furthermore, the server determines the simulation scenario test score result based on the difference between the target vehicle's projection distance and the autonomous driving vehicle's projection distance (max_L2-max_L1 or max_L1-max_L2), and the difference between the time when the target vehicle starts changing lanes and the time when the autonomous driving vehicle changes lanes (t1-t2 or t2-t1).

[0049] It should be noted that if the autonomous vehicle's projected distance is greater than the target vehicle's projected distance, that is, the autonomous vehicle occupies a larger span of the target obstacle vehicle's frontal area than the target vehicle does, then the simulation scenario test score will be higher. If the autonomous vehicle starts changing lanes earlier than the vehicle changes lanes, then the simulation scenario test score will be higher. In other words, the earlier the autonomous vehicle takes over the target obstacle vehicle's frontal area than the target vehicle, the higher the simulation scenario test score will be. Accordingly, the simulation scenario test score is linearly correlated with the difference between the target vehicle's projected distance and the autonomous vehicle's projected distance, and with the difference between the target vehicle's lane change initiation time and the autonomous vehicle's lane change time.

[0050] In an embodiment of the present invention, an autonomous driving behavior test is performed using information from multiple actual crowded lane change scenarios and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. A lane change behavior similarity scoring process is performed based on the information from multiple actual crowded lane change scenarios and the information from multiple autonomous driving vehicle lane change simulation test information to obtain a simulation scenario test scoring result, thereby improving the test scoring accuracy and efficiency of the autonomous driving software in handling crowded lane change scenarios and improving the safety of autonomous driving.

[0051] See also Figure 3 Another embodiment of the method for evaluating a crowded lane change simulation scenario in the embodiment of the present invention includes:

[0052] 301. Acquire actual vehicle driving scene information corresponding to a target vehicle, where the actual vehicle driving scene information includes position information, speed information, turn signal flashing information, and obstacle information around the target vehicle.

[0053] The specific execution process of step 301 is similar to the specific execution process of step 101, and will not be repeated here.

[0054] 302. Perform congested lane change scene segmentation on the actual vehicle driving scene information to obtain multiple pieces of congested lane change actual scene information.

[0055] The specific execution process of step 302 is similar to the specific execution process of step 102, and will not be repeated here.

[0056] 303. Perform an autonomous driving behavior test based on multiple actual crowded lane change scenario information and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario.

[0057] Among them, the actual scene information of each crowded lane change corresponds one-to-one to each crowded lane change simulation scene, and the turn signal flashing information includes the crowded lane change time and the corresponding crowded lane change direction corresponding to each crowded lane change simulation scene. The crowded lane change time corresponding to each crowded lane change simulation scene is used to indicate the turn signal turning on time, and the crowded lane change direction corresponding to each crowded lane change simulation scene is used to indicate the turn signal direction.

[0058] In some embodiments, the server runs the autonomous driving simulation algorithm to be tested based on multiple actual crowded lane change scenario information, obtaining multiple crowded lane change simulation scenarios. That is, the server runs the autonomous driving simulation algorithm to be tested under different actual crowded lane change scenario information to obtain the crowded lane change simulation scenario corresponding to each actual crowded lane change scenario information. The server reads the crowded lane change time and the corresponding crowded lane change direction corresponding to each crowded lane change simulation scenario from the turn signal flashing information. In each crowded lane change simulation scenario, the server sends a lane change request to the autonomous driving vehicle in the crowded lane change direction according to the crowded lane change time corresponding to each crowded lane change simulation scenario, obtaining autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. The lane change request in the crowded lane change direction is used to indicate that the lane change direction of the autonomous driving vehicle in each crowded lane change simulation scenario is consistent with the lane change direction of the target vehicle, that is, the autonomous driving vehicle and the target vehicle are changing lanes in the same direction.

[0059] 304. Perform lane change behavior similarity scoring based on multiple actual crowded lane change scenario information and autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario to obtain simulation scenario test scoring results. The lane change behavior includes the lane change trajectory and the lane change start time.

[0060] That is, the server performs lane change trajectory similarity scoring and lane change start time similarity scoring on the lane change simulation test information of the autonomous driving vehicle corresponding to each crowded lane change simulation scenario based on multiple actual crowded lane change scene information. In some embodiments, the server obtains target vehicle lane change information corresponding to each congested lane change simulation scenario from multiple actual congested lane change scenario information. The server determines the target vehicle lane change start time and target vehicle projection distance corresponding to each congested lane change simulation scenario based on the target vehicle lane change information corresponding to each congested lane change simulation scenario. The target vehicle projection distance indicates the vehicle body length of the target vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed. The server determines the autonomous driving vehicle lane change start time and autonomous driving vehicle projection distance corresponding to each congested lane change simulation scenario based on autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario. The autonomous driving vehicle projection distance indicates the vehicle body length of the autonomous driving vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed. It is understandable that there is a one-to-one correspondence between the target obstacle vehicle in each actual congested lane change scenario information and the target obstacle vehicle in each congested lane change simulation scenario. In this embodiment, they are all referred to as target obstacle vehicles. The server performs lane change behavior similarity scoring based on the target vehicle lane change start time, target vehicle projection distance, autonomous driving vehicle lane change start time, and autonomous driving vehicle projection distance to obtain a simulation scenario test scoring result. The target vehicle projection distance and the autonomous driving vehicle projection distance may be equal or different. The target vehicle lane change start time and the autonomous driving vehicle lane change start time may be the same or different. Figure 2 As shown in the figure, in each actual crowded lane change scenario information, AB is the body length of the target vehicle a projected to the target obstacle vehicle b along the direction perpendicular to the lane line during the crowded lane change process; in each crowded lane change simulation scenario, AB is the body length of the autonomous driving vehicle a projected to the target obstacle vehicle b along the direction perpendicular to the lane line during the crowded lane change process.

[0061] Furthermore, when the server performs a lane change behavior similarity scoring process based on the target vehicle lane change start time, the target vehicle projection distance, the autonomous driving vehicle lane change start time and the autonomous driving vehicle projection distance to obtain a simulation scenario test scoring result step, in some embodiments, the server calculates the projection distance difference corresponding to each congested lane change simulation scenario according to the target vehicle projection distance and the autonomous driving vehicle projection distance, and obtains the lane change completion score corresponding to each congested lane change simulation scenario based on the projection distance difference corresponding to each congested lane change simulation scenario, wherein the projection distance difference corresponding to each congested lane change simulation scenario is the difference (max_L2-max_L1) between the autonomous driving vehicle projection distance max_L2 and the target vehicle projection distance max_L1, and the relationship between the lane change completion score complete_score corresponding to each congested lane change simulation scenario and max_L1 and max_L2 is as follows: Figure 4 As shown, the functional relationship is The server calculates the lane change time difference corresponding to each crowded lane change simulation scenario according to the lane change start time of the target vehicle and the lane change start time of the autonomous driving vehicle, and obtains the lane change time score corresponding to each crowded lane change simulation scenario based on the lane change time difference corresponding to each crowded lane change simulation scenario. Among them, the lane change time difference corresponding to each crowded lane change simulation scenario is the difference (t2-t1) between the lane change start time t2 of the autonomous driving vehicle and the lane change start time t1 of the target vehicle. The relationship between the lane change time score start_time_score corresponding to each crowded lane change simulation scenario and t1 and t2 is as follows Figure 5 As shown, the functional relationship is The server sets the crowded lane change scene score corresponding to each crowded lane change simulation scenario based on the lane change completion score and the lane change time score corresponding to each crowded lane change simulation scenario, where score = complete_score × start_time_score; the server summarizes the crowded lane change scene scores corresponding to each crowded lane change simulation scenario to obtain the simulation scenario test score results.

[0062] like Figure 2As shown, for example, in the crowded lane change actual scene information S1, the body width of the target vehicle a and the body width of the target obstacle vehicle b are both 2 meters. Correspondingly, in the crowded lane change simulation scene S1', the body width of the autonomous driving vehicle a and the body width of the target obstacle vehicle b are also 2 meters. When both the crowded lane change actual scene information S1 and the crowded lane change simulation scene S1' have completed the crowded lane change, and the target vehicle and the autonomous driving vehicle are both located directly in front of the target obstacle vehicle, the target vehicle projection distance max_L1 and the autonomous driving vehicle projection distance max_L2 are both 2 meters. The server then calculates the projection distance difference max_L2-max_L1 corresponding to the crowded lane change simulation scene as 0 meters, as shown in FIG. Figure 4 As shown, the server calculates the lane change completion score corresponding to each crowded lane change simulation scenario as 1.0 based on the projection distance difference of 0 meters. In the crowded lane change actual scene information S1, the target vehicle a begins to occupy the position in front of the target obstacle vehicle b at 7.5 seconds after the target vehicle lane change start time. In the crowded lane change simulation scene S1', the autonomous driving vehicle a starts changing lanes later than the target vehicle a. The autonomous driving vehicle a begins to occupy the position in front of the target obstacle vehicle b at 9.6 seconds. Therefore, the target vehicle lane change start time t1 is 7.5 seconds, and the autonomous driving vehicle lane change start time t2 is 9.6 seconds. The lane change time difference t2-t1 is 2.1 seconds. Figure 5 As shown, the server calculates the lane change time score start_time_score as 0.42 based on the lane change time difference of 2.1 seconds. The server then calculates the crowded lane change scenario score as 0.42×1.0, which is 0.42. In this embodiment, the simulation scenario test score is positively correlated with the difference between the target vehicle's projected distance and the autonomous vehicle's projected distance (i.e., max_L2 - max_L1), and negatively correlated with the difference between the target vehicle's lane change start time and the autonomous vehicle's lane change time (i.e., t2 - t1). Furthermore, the lane change completion score complete_score and the lane change time score start_time_score corresponding to each crowded lane change simulation scenario each have a value range greater than or equal to 0 and less than or equal to 1.

[0063] It is understood that the server calculates a corresponding crowded lane change scenario score for each crowded lane change simulation scenario, and then obtains a simulation scenario test score result based on the sum of the crowded lane change scenario scores corresponding to each crowded lane change simulation scenario. Furthermore, the server may also calculate an average crowded lane change simulation scenario score based on the crowded lane change scenario scores corresponding to each crowded lane change simulation scenario, count the number of crowded lane change simulation scenario scores that exceed a preset threshold, and generate a crowded lane change simulation scenario score histogram. The server then uses the average crowded lane change simulation scenario score, the number of crowded lane change simulation scenario scores that exceed the preset threshold, and / or the crowded lane change simulation scenario score histogram as the simulation scenario test score result, thereby evaluating the level of the autonomous driving algorithm to be tested in handling crowded lane change scenarios.

[0064] 305. Determine whether the simulation scenario test scoring result meets a preset scoring threshold.

[0065] Specifically, the server calculates the difference between the simulation scenario test score result and a preset score threshold to obtain a score difference. If the score difference is greater than or equal to the preset score, the server determines that the simulation scenario test score result meets the preset score threshold. If the score difference is less than the preset score, the server determines that the simulation scenario test score result does not meet the preset score threshold. The preset score can be 0 or other values, which are not specifically limited here.

[0066] 306. If the simulation scenario test score result meets the preset score threshold, a test pass label is set for the autonomous driving simulation algorithm to be tested, and a version release process is performed on the autonomous driving simulation algorithm to be tested.

[0067] Specifically, if the simulation scenario test score meets a preset score threshold, the server sets a test pass label for the autonomous driving simulation algorithm to be tested according to preset label setting rules. Based on the test pass label, the server sets the test completion progress of the version of the autonomous driving simulation algorithm to be tested as completed. The server generates software version release information for the autonomous driving simulation algorithm to be tested, packages the autonomous driving simulation algorithm to be tested and the software version release information to obtain the autonomous driving simulation algorithm to be released, and calls a preset version release review interface to perform a version release review on the autonomous driving simulation algorithm to be released. Furthermore, the server tests different versions of the autonomous driving algorithm using multiple real-world crowded lane change scenarios, thereby eliminating autonomous driving test errors caused by different crowded lane change scenarios using the same real-world crowded lane change scenario information, thereby improving autonomous driving test accuracy.

[0068] In an embodiment of the present invention, an autonomous driving behavior test is performed using information from multiple actual crowded lane change scenarios and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. A lane change behavior similarity scoring process is performed based on the information from multiple actual crowded lane change scenarios and the information from multiple autonomous driving vehicle lane change simulation test information to obtain a simulation scenario test scoring result, thereby improving the test scoring accuracy and efficiency of the autonomous driving software in handling crowded lane change scenarios and improving the safety of autonomous driving.

[0069] The above describes the crowded lane change simulation scenario evaluation method according to the embodiment of the present invention. The following describes the crowded lane change simulation scenario evaluation device according to the embodiment of the present invention. Figure 6 An embodiment of a crowded lane change simulation scenario evaluation device according to an embodiment of the present invention includes:

[0070] The first acquisition module 601 is used to obtain actual vehicle driving scene information corresponding to the target vehicle, wherein the actual vehicle driving scene information includes the position information, speed information, turn signal flashing information of the target vehicle, and obstacle information around the target vehicle;

[0071] A segmentation module 602 is configured to segment the actual vehicle driving scene information into congested lane change scenes to obtain a plurality of congested lane change actual scene information;

[0072] A testing module 603 is configured to perform an autonomous driving behavior test based on the multiple congested lane change actual scenario information and the turn signal flashing information, and obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario;

[0073] Scoring module 604 is configured to perform lane change behavior similarity scoring based on the multiple actual crowded lane change scenario information and the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario, thereby obtaining a simulation scenario test scoring result. The lane change behavior includes the lane change trajectory and the lane change start time.

[0074] In an embodiment of the present invention, an autonomous driving behavior test is performed using information from multiple actual crowded lane change scenarios and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. A lane change behavior similarity scoring process is performed based on the information from multiple actual crowded lane change scenarios and the information from multiple autonomous driving vehicle lane change simulation test information to obtain a simulation scenario test scoring result, thereby improving the test scoring accuracy and efficiency of the autonomous driving software in handling crowded lane change scenarios and improving the safety of autonomous driving.

[0075] See also Figure 7 Another embodiment of the crowded lane change simulation scenario evaluation device according to the embodiment of the present invention includes:

[0076] The first acquisition module 601 is used to obtain actual vehicle driving scene information corresponding to the target vehicle, wherein the actual vehicle driving scene information includes the position information, speed information, turn signal flashing information of the target vehicle, and obstacle information around the target vehicle;

[0077] A segmentation module 602 is configured to segment the actual vehicle driving scene information into congested lane change scenes to obtain a plurality of congested lane change actual scene information;

[0078] A testing module 603 is configured to perform an autonomous driving behavior test based on the multiple congested lane change actual scenario information and the turn signal flashing information, and obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario;

[0079] Scoring module 604 is configured to perform lane change behavior similarity scoring based on the multiple actual crowded lane change scenario information and the autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario, thereby obtaining a simulation scenario test scoring result. The lane change behavior includes the lane change trajectory and the lane change start time.

[0080] In a feasible implementation manner, the segmentation module 602 is specifically configured to:

[0081] Performing scene segmentation processing on the actual vehicle driving scene information using a preset scene segmentation model to obtain a scene segmentation result;

[0082] The scene segmentation results are subjected to scene filtering according to preset congested lane change scene screening rules to obtain multiple congested lane change actual scene information.

[0083] In a feasible implementation manner, the testing module 603 is specifically configured to:

[0084] Running the autonomous driving simulation algorithm to be tested according to the multiple congested lane change actual scene information respectively to obtain multiple congested lane change simulation scenes;

[0085] Reading the congested lane change time and the congested lane change direction corresponding to each congested lane change simulation scenario from the turn signal flashing information;

[0086] In each congested lane change simulation scenario, a lane change request in the congested lane change direction is sent to the autonomous driving vehicle according to the congested lane change time corresponding to each congested lane change simulation scenario, and autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario is obtained.

[0087] In a feasible implementation, the scoring module 604 further includes:

[0088] An acquiring unit 6041 is configured to acquire lane change information of a target vehicle corresponding to each congested lane change simulation scenario from the plurality of congested lane change actual scenario information;

[0089] A first determining unit 6042 is configured to determine, based on the lane change information of the target vehicle corresponding to each congested lane change simulation scenario, a lane change start time and a target vehicle projection distance corresponding to each congested lane change simulation scenario. The target vehicle projection distance indicates the length of the target vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed.

[0090] A second determining unit 6043 is configured to determine, based on the lane change simulation test information of the autonomous driving vehicle corresponding to each congested lane change simulation scenario, a lane change start time and a projection distance of the autonomous driving vehicle corresponding to each congested lane change simulation scenario, wherein the projection distance indicates the length of the autonomous driving vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed;

[0091] The scoring unit 6044 is used to perform lane change behavior similarity scoring processing based on the target vehicle's lane change start time, the target vehicle's projected distance, the autonomous driving vehicle's lane change start time, and the autonomous driving vehicle's projected distance to obtain a simulation scenario test scoring result.

[0092] In a feasible implementation, the scoring unit 6044 is specifically configured to:

[0093] Calculating a projection distance difference corresponding to each crowded lane change simulation scenario according to the target vehicle projection distance and the autonomous driving vehicle projection distance, and obtaining a lane change completion score corresponding to each crowded lane change simulation scenario according to the projection distance difference corresponding to each crowded lane change simulation scenario;

[0094] Calculating a lane change time difference corresponding to each congested lane change simulation scenario according to the lane change start time of the target vehicle and the lane change start time of the autonomous driving vehicle, and obtaining a lane change time score corresponding to each congested lane change simulation scenario according to the lane change time difference corresponding to each congested lane change simulation scenario;

[0095] Based on the lane change completion score and the lane change time score corresponding to each crowded lane change simulation scenario, a crowded lane change scenario score corresponding to each crowded lane change simulation scenario is set;

[0096] The crowded lane-changing scenario scores corresponding to each crowded lane-changing simulation scenario are summarized to obtain the simulation scenario test score results.

[0097] In a feasible implementation manner, the crowded lane change simulation scenario evaluation device further includes:

[0098] A second acquisition module 605 is configured to acquire initial vehicle driving scene information in an actual road driving scenario, wherein the initial vehicle driving scene information includes vehicle perception information and vehicle video image information;

[0099] A verification module 606 is configured to perform information verification processing on the initial vehicle driving scene information to obtain a verification result;

[0100] The labeling module 607 is used to perform data analysis, labeling and data packaging processing on the initial vehicle driving scene information when the verification result is verification passed, so as to obtain the actual vehicle driving scene information corresponding to the target vehicle.

[0101] In a feasible implementation manner, the crowded lane change simulation scenario evaluation device further includes:

[0102] A judgment module 608 is used to judge whether the simulation scenario test score result meets a preset score threshold;

[0103] The setting module 609 is used to set a test pass label for the autonomous driving simulation algorithm to be tested if the simulation scenario test score result meets the preset score threshold, and perform version release processing on the autonomous driving simulation algorithm to be tested.

[0104] In an embodiment of the present invention, an autonomous driving behavior test is performed using information from multiple actual crowded lane change scenarios and turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each crowded lane change simulation scenario. A lane change behavior similarity scoring process is performed based on the information from multiple actual crowded lane change scenarios and the information from multiple autonomous driving vehicle lane change simulation test information to obtain a simulation scenario test scoring result, thereby improving the test scoring accuracy and efficiency of the autonomous driving software in handling crowded lane change scenarios and improving the safety of autonomous driving.

[0105] above Figure 6 and Figure 7 The crowded lane change simulation scenario evaluation device in the embodiment of the present invention is described in detail from a modular perspective. The crowded lane change simulation scenario evaluation device in the embodiment of the present invention is described in detail from a hardware processing perspective.

[0106] Figure 8: This is a structural diagram of a crowded lane change simulation scenario evaluation device provided by an embodiment of the present invention. The crowded lane change simulation scenario evaluation device 800 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 810 (for example, one or more processors) and a memory 820, and one or more storage media 830 (for example, one or more mass storage devices) storing application programs 833 or data 832. Among them, the memory 820 and the storage medium 830 can be temporary storage or permanent storage. The program stored in the storage medium 830 may include one or more modules (not shown in the figure), and each module may include a series of computer program operations in the crowded lane change simulation scenario evaluation device 800. Furthermore, the processor 810 can be configured to communicate with the storage medium 830 to execute a series of computer program operations in the storage medium 830 on the crowded lane change simulation scenario evaluation device 800.

[0107] The crowded lane change simulation scenario evaluation device 800 may further include one or more power supplies 840, one or more wired or wireless network interfaces 850, one or more input and output interfaces 860, and / or one or more operating systems 831, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 8 The structure of the crowded lane-changing simulation scenario evaluation device shown does not constitute a limitation on the crowded lane-changing simulation scenario evaluation device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0108] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, cause the computer to execute the steps of the crowded lane change simulation scenario evaluation method.

[0109] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program code.

[0111] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. 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 invention.

Claims

1. A method for evaluating a crowded lane-changing simulation scenario, characterized in that: The congested lane change simulation scenario evaluation method includes: Acquire actual vehicle driving scene information corresponding to the target vehicle, the actual vehicle driving scene information including the position information, speed information, turn signal flashing information of the target vehicle, and obstacle information around the target vehicle; Performing congested lane-changing scene segmentation on the actual vehicle driving scene information to obtain a plurality of congested lane-changing actual scene information; Performing an autonomous driving behavior test based on the multiple congested lane change actual scene information and the turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene; performing lane change behavior similarity scoring based on the multiple actual crowded lane change scenario information and the lane change simulation test information of the autonomous driving vehicle corresponding to each crowded lane change simulation scenario to obtain a simulation scenario test scoring result, wherein the lane change behavior includes a lane change trajectory and a lane change start time; The lane change behavior similarity scoring process is performed based on the multiple congested lane change actual scene information and the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene to obtain a simulation scene test scoring result, including: Acquiring target vehicle lane change information corresponding to each congested lane change simulation scenario from the plurality of congested lane change actual scenario information; Determine the lane change start time and target vehicle projection distance corresponding to each congested lane change simulation scenario based on the target vehicle lane change information corresponding to each congested lane change simulation scenario. The target vehicle projection distance indicates the length of the target vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed. Determining the lane change start time and the projection distance of the autonomous vehicle corresponding to each congested lane change simulation scenario based on the autonomous vehicle lane change simulation test information corresponding to each congested lane change simulation scenario, wherein the projection distance indicates the length of the autonomous vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed; A lane-changing behavior similarity scoring process is performed based on the target vehicle's lane-changing start time, the target vehicle's projected distance, the autonomous driving vehicle's lane-changing start time, and the autonomous driving vehicle's projected distance to obtain a simulation scenario test scoring result.

2. The crowded lane change simulation scenario evaluation method according to claim 1, characterized in that: The congested lane change scene segmentation is performed on the actual vehicle driving scene information to obtain a plurality of congested lane change actual scene information, including: Performing scene segmentation processing on the actual vehicle driving scene information using a preset scene segmentation model to obtain a scene segmentation result; The scene segmentation results are subjected to scene filtering according to preset congested lane change scene screening rules to obtain multiple congested lane change actual scene information.

3. The crowded lane change simulation scenario evaluation method according to claim 1, characterized in that: The autonomous driving behavior test is performed based on the multiple congested lane change actual scene information and the turn signal flashing information to obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene, including: Running the autonomous driving simulation algorithm to be tested according to the multiple congested lane change actual scene information respectively to obtain multiple congested lane change simulation scenes; Reading the congested lane change time and the congested lane change direction corresponding to each congested lane change simulation scenario from the turn signal flashing information; In each congested lane change simulation scenario, a lane change request in the congested lane change direction is sent to the autonomous driving vehicle according to the congested lane change time corresponding to each congested lane change simulation scenario, and autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario is obtained.

4. The method for evaluating a crowded lane change simulation scenario according to claim 1, wherein: The lane change behavior similarity scoring process is performed based on the target vehicle lane change start time, the target vehicle projection distance, the autonomous driving vehicle lane change start time, and the autonomous driving vehicle projection distance to obtain a simulation scenario test scoring result, including: Calculating a projection distance difference corresponding to each crowded lane change simulation scenario according to the target vehicle projection distance and the autonomous driving vehicle projection distance, and obtaining a lane change completion score corresponding to each crowded lane change simulation scenario according to the projection distance difference corresponding to each crowded lane change simulation scenario; Calculating a lane change time difference corresponding to each congested lane change simulation scenario according to the lane change start time of the target vehicle and the lane change start time of the autonomous driving vehicle, and obtaining a lane change time score corresponding to each congested lane change simulation scenario according to the lane change time difference corresponding to each congested lane change simulation scenario; Based on the lane change completion score and the lane change time score corresponding to each crowded lane change simulation scenario, a crowded lane change scenario score corresponding to each crowded lane change simulation scenario is set; The crowded lane-changing scenario scores corresponding to each crowded lane-changing simulation scenario are summarized to obtain the simulation scenario test score results.

5. The method for evaluating a crowded lane change simulation scenario according to any one of claims 1 to 4, characterized in that: Before obtaining actual vehicle driving scene information corresponding to the target vehicle, wherein the actual vehicle driving scene information includes position information, speed information, turn signal flashing information, and obstacle information around the target vehicle, the congested lane change simulation scene evaluation method further includes: Acquiring initial vehicle driving scene information in an actual road driving scene, the initial vehicle driving scene information including vehicle perception information and vehicle video image information; Performing information verification processing on the initial vehicle driving scene information to obtain a verification result; When the verification result is that the verification is passed, data analysis, annotation and data packaging processing are performed on the initial vehicle driving scene information to obtain actual vehicle driving scene information corresponding to the target vehicle.

6. The method for evaluating a crowded lane change simulation scenario according to any one of claims 1 to 4, characterized in that: After performing lane change behavior similarity scoring processing based on the multiple congested lane change actual scenario information and the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario to obtain a simulation scenario test scoring result, wherein the lane change behavior includes a lane change trajectory and a lane change start time, the congested lane change simulation scenario evaluation method further includes: Determine whether the simulation scenario test score meets a preset score threshold; If the simulation scenario test scoring result meets the preset scoring threshold, a test pass label is set for the autonomous driving simulation algorithm to be tested, and a version release process is performed on the autonomous driving simulation algorithm to be tested.

7. A crowded lane change simulation scenario evaluation device, characterized in that: The crowded lane change simulation scenario evaluation device comprises: A first acquisition module is used to obtain actual vehicle driving scene information corresponding to the target vehicle, wherein the actual vehicle driving scene information includes the position information, speed information, turn signal flashing information of the target vehicle, and obstacle information around the target vehicle; a segmentation module, configured to segment the actual vehicle driving scene information into a congested lane change scene to obtain a plurality of congested lane change actual scene information; a testing module, configured to perform an autonomous driving behavior test based on the multiple congested lane change actual scenario information and the turn signal flashing information, and obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario; a scoring module configured to perform lane change behavior similarity scoring based on the multiple actual crowded lane change scenario information and the lane change simulation test information of the autonomous driving vehicle corresponding to each crowded lane change simulation scenario, thereby obtaining a simulation scenario test scoring result, wherein the lane change behavior includes a lane change trajectory and a lane change start time; The scoring module further includes: an acquisition unit for acquiring target vehicle lane change information corresponding to each congested lane change simulation scenario from the multiple congested lane change actual scenario information; a first determination unit for determining, based on the target vehicle lane change information corresponding to each congested lane change simulation scenario, a target vehicle lane change start time and a target vehicle projection distance corresponding to each congested lane change simulation scenario, the target vehicle projection distance being used to indicate the vehicle body length of the target vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed; a second determination unit for determining, based on the autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scenario, a lane change start time and an autonomous driving vehicle projection distance corresponding to each congested lane change simulation scenario, the autonomous driving vehicle projection distance being used to indicate the vehicle body length of the autonomous driving vehicle projected perpendicular to the lane line to the target obstacle vehicle when the congested lane change is completed; and a scoring unit for performing lane change behavior similarity scoring processing based on the target vehicle lane change start time, the target vehicle projection distance, the autonomous driving vehicle lane change start time, and the autonomous driving vehicle projection distance to obtain a simulation scenario test scoring result.

8. The crowded lane change simulation scenario evaluation device according to claim 7, characterized in that: The test module is specifically used to: run the autonomous driving simulation algorithm to be tested according to the multiple congested lane change actual scene information respectively to obtain multiple congested lane change simulation scenes; read the congested lane change time and the corresponding congested lane change direction corresponding to each congested lane change simulation scene from the turn signal flashing information; in each congested lane change simulation scene, send a lane change request to the autonomous driving vehicle in the congested lane change direction according to the congested lane change time corresponding to each congested lane change simulation scene, and obtain autonomous driving vehicle lane change simulation test information corresponding to each congested lane change simulation scene.

9. A crowded lane change simulation scenario evaluation device, characterized in that: The crowded lane change simulation scenario evaluation device includes: a memory and at least one processor, wherein the memory stores a computer program; The at least one processor calls the computer program in the memory to enable the crowded lane change simulation scenario evaluation device to execute the crowded lane change simulation scenario evaluation method according to any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for evaluating a crowded lane change simulation scenario as described in any one of claims 1 to 6 is implemented.

Images