Congested merging scenario assessment method, device, equipment and storage medium

Abstract

The present invention relates to the field of autonomous driving technology, and discloses a method, apparatus, device, and storage medium for evaluating congested merging scenarios, which are used to improve the accuracy of evaluating autonomous driving software in handling congested merging scenarios and improve the efficiency of autonomous driving testing. The congested merging scenario evaluation method comprises: obtaining a vehicle driving scene dataset corresponding to a target vehicle, the vehicle driving scene dataset being used to indicate scene data collected by the target vehicle during driving in an actual road scene; performing data cleaning and data framing processing on the vehicle driving scene dataset in sequence according to a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; performing merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested merging scene data; performing scene simulation and autonomous driving behavior evaluation processing on the target congested merging scene data to obtain an evaluation result.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method, device, equipment and storage medium for evaluating a crowded merging scenario. Background Art

[0002] A lane merge occurs when a vehicle merges left or right from its current lane into another lane where there is an intersection. For example, a three-lane road becomes a two-lane intersection, with vehicles in both lanes simultaneously moving into a single lane. Congested merging scenarios are one of the most challenging for autonomous driving software to handle.

[0003] In the existing technology, the evaluation method for autonomous driving software's handling of congested lane-merging 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 congested lane-merging scenarios in actual road testing, the scenario evaluation accuracy is poor. Summary of the Invention

[0004] The present invention provides a method, apparatus, device, and storage medium for evaluating crowded merging scenarios, which are used to improve the accuracy of evaluating autonomous driving software's handling of crowded merging scenarios, improve the efficiency of autonomous driving testing, and improve the applicability and 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 merging scene, comprising: obtaining a vehicle driving scene dataset corresponding to a target vehicle, the vehicle driving scene dataset being used to indicate scene data collected during the driving of the target vehicle in an actual road scene; performing data cleaning and data framing processing on the vehicle driving scene dataset in sequence through a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; performing merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target crowded merging scene data; performing scene simulation and autonomous driving behavior evaluation processing on the target crowded merging scene data to obtain an evaluation result.

[0006] In a feasible implementation, the method of performing lane merging scene recognition and vehicle congestion detection on the multiple framed scene data according to a preset scene screening algorithm to obtain target congested lane merging scene data includes: performing vehicle driving trajectory analysis on each framed scene data according to the preset scene screening algorithm to obtain a lane position relationship corresponding to each framed scene data, wherein the lane position relationship is used to indicate relative vehicle lane position information between the target vehicle and a reference obstacle vehicle; judging whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle according to the lane position relationship corresponding to each framed scene data, wherein the target lane merging behavior includes a left lane merging behavior and a right lane merging behavior; if there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle, generating the target congested lane merging scene data based on each framed scene data corresponding to the target lane merging behavior.

[0007] In one feasible implementation, the determining whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle according to the lane position relationship corresponding to each framed scene data includes: determining whether the target vehicle and the reference obstacle vehicle meet a preset merging scenario correlation rule according to the lane position relationship corresponding to each framed scene data; if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, determining whether the lane in which the reference obstacle vehicle was located before merging was a congested lane; if the lane in which the reference obstacle vehicle was located before merging was a congested lane, determining that there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle.

[0008] In a feasible implementation manner, judging whether the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule according to the lane position relationship corresponding to each framed scene data includes: obtaining, according to the lane position relationship corresponding to each framed scene data, a first moment corresponding to when the target vehicle travels to the target position, a second moment corresponding to when the reference obstacle vehicle travels to the target position, and a target distance difference, the target distance difference being used to indicate the minimum distance between the target vehicle and the reference obstacle vehicle before reaching the target position; calculating the target moment difference based on the first moment and the second moment; and determining that the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule when the target moment difference is less than a preset time threshold and the target distance difference is less than a preset distance difference.

[0009] In one feasible implementation, if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, then determining whether the lane in which the reference obstacle vehicle was located before merging is a congested lane includes: if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, obtaining the number of other obstacle vehicles within a preset distance of the target lane, where the target lane is the lane in which the reference obstacle vehicle was located before merging; determining whether the number of other obstacle vehicles is greater than a preset number threshold; if the number of other obstacle vehicles is greater than the preset number threshold, determining that the lane in which the reference obstacle vehicle was located before merging is a congested lane.

[0010] In a feasible implementation, the performing of scene simulation and autonomous driving behavior evaluation processing based on the target congested merging scene data to obtain an evaluation result includes: performing a merging scene simulation on the target congested merging scene data through preset autonomous driving simulation software to obtain autonomous driving congested merging simulation scene data; triggering the autonomous driving vehicle simulated driving in the autonomous driving congested merging simulation scene data according to preset control instructions to obtain congested merging simulation behavior data corresponding to the autonomous driving vehicle; extracting the actual congested merging behavior data corresponding to the target vehicle from the target congested merging scene data, and performing scene evaluation processing on the congested merging simulation behavior data and the actual congested merging behavior data according to preset scene scoring rules to obtain an evaluation result.

[0011] In a feasible implementation, after performing scenario simulation and autonomous driving behavior evaluation processing based on the target congested merging scenario data to obtain an evaluation result, the congested merging scenario evaluation method further includes: generating a congested merging scenario evaluation report based on preset autonomous driving simulation software, the target congested merging scenario data and the evaluation result; and pushing the congested merging scenario evaluation report to a target terminal so that the target terminal performs a visual chart display according to the congested merging scenario evaluation report.

[0012] The second aspect of the present invention provides a congested merging scene evaluation device, comprising: an acquisition module for acquiring a vehicle driving scene data set corresponding to a target vehicle, wherein the vehicle driving scene data set is used to indicate scene data collected during the driving of the target vehicle in an actual road scene; a processing module for performing data cleaning and data framing processing on the vehicle driving scene data set in sequence through a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; a screening module for performing merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested merging scene data; an evaluation module for performing scene simulation and automatic driving behavior evaluation processing based on the target congested merging scene data to obtain an evaluation result.

[0013] In a feasible embodiment, the screening module further includes: an analysis unit, configured to perform vehicle driving trajectory analysis on each framed scene data according to a preset scene screening algorithm to obtain a lane position relationship corresponding to each framed scene data, wherein the lane position relationship is used to indicate relative vehicle lane position information between the target vehicle and the reference obstacle vehicle; a judgment unit, configured to judge whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle according to the lane position relationship corresponding to each framed scene data, wherein the target lane merging behavior includes a left lane merging behavior and a right lane merging behavior; and a generation unit, configured to generate target congested lane merging scene data based on each framed scene data corresponding to the target lane merging behavior if there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle.

[0014] In a feasible embodiment, the judgment unit further includes: a first judgment subunit, configured to judge whether the target vehicle and the reference obstacle vehicle meet a preset merging scenario correlation rule based on the lane position relationship corresponding to each framed scene data; a second judgment subunit, configured to judge whether the lane in which the reference obstacle vehicle was located before merging is a congested lane if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule; and a determination subunit, configured to determine whether a target merging behavior exists between the target vehicle and the reference obstacle vehicle if the lane in which the reference obstacle vehicle was located before merging is a congested lane.

[0015] In a feasible implementation manner, the first judgment subunit is specifically used to: obtain, according to the lane position relationship corresponding to each framed scene data, a first moment corresponding to when the target vehicle travels to the target position, a second moment corresponding to when the reference obstacle vehicle travels to the target position, and a target distance difference, wherein the target distance difference is used to indicate the minimum distance between the target vehicle and the reference obstacle vehicle before reaching the target position; calculate the target moment difference based on the first moment and the second moment; when the target moment difference is less than a preset time threshold and the target distance difference is less than a preset distance difference, determine that the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rule.

[0016] In one feasible embodiment, the second judgment subunit is specifically configured to: if the target vehicle and the reference obstacle vehicle meet a preset merging scenario correlation rule, obtain the number of other obstacle vehicles within a preset distance of a target lane, where the target lane is the lane in which the reference obstacle vehicle was located before merging; determine whether the number of other obstacle vehicles is greater than a preset number threshold; and if the number of other obstacle vehicles is greater than the preset number threshold, determine that the lane in which the reference obstacle vehicle was located before merging is a congested lane.

[0017] In a feasible implementation, the evaluation module is specifically used to: perform a merging scenario simulation on the target crowded merging scenario data through preset autonomous driving simulation software to obtain autonomous driving crowded merging simulation scenario data; trigger the simulated driving of the autonomous driving vehicle in the autonomous driving crowded merging simulation scenario data according to preset control instructions to obtain crowded merging simulation behavior data corresponding to the autonomous driving vehicle; extract the crowded merging actual behavior data corresponding to the target vehicle from the target crowded merging scene data, and perform scenario evaluation processing on the crowded merging simulation behavior data and the crowded merging actual behavior data according to preset scenario scoring rules to obtain an evaluation result.

[0018] In a feasible implementation, the congested merging scenario assessment device further includes: a generation module for generating a congested merging scenario assessment report based on preset autonomous driving simulation software, the target congested merging scenario data and the assessment results; and a push module for pushing the congested merging scenario assessment report to a target terminal, so that the target terminal performs a visual chart display according to the congested merging scenario assessment report.

[0019] A third aspect of the present invention provides a crowded merging scene assessment device, comprising: a memory and at least one processor, wherein the memory stores instructions, 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 merging scene assessment device executes the above-mentioned crowded merging scene assessment 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 congested merging scenario.

[0021] In the technical solution provided by the present invention, a vehicle driving scene dataset corresponding to a target vehicle is obtained, wherein the vehicle driving scene dataset is used to indicate scene data collected during the target vehicle's driving in an actual road scene; data cleaning and data framing processing are sequentially performed on the vehicle driving scene dataset using a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; merging scene recognition and vehicle congestion detection are performed on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested merging scene data; and scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain an evaluation result. In an embodiment of the present invention, target congested merging scene data is extracted from the vehicle driving scene dataset using a preset cleaning strategy, preset framing rules, and preset scene screening algorithm, thereby improving the accuracy of target congested merging scene data screening; scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain an evaluation result, thereby improving the accuracy of evaluating the autonomous driving software's handling of congested merging scenes, improving the efficiency of autonomous driving testing, and improving the applicability and safety of autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of an embodiment of a method for evaluating a congested merging scenario according to an embodiment of the present invention;

[0023] Figure 2 A schematic diagram of an embodiment of screening target congestion and merging scene data in an embodiment of the present invention;

[0024] Figure 3 A schematic diagram of another embodiment of a method for evaluating a congested merging scenario according to an embodiment of the present invention;

[0025] Figure 4 A schematic diagram of an embodiment of a device for evaluating a congested merging scenario according to an embodiment of the present invention;

[0026] Figure 5 2 is a schematic diagram of another embodiment of a device for evaluating a congested merging scenario according to an embodiment of the present invention;

[0027] Figure 6 FIG. 1 is a schematic diagram of an embodiment of a device for evaluating a congested merging scenario according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] Embodiments of the present invention provide a method, apparatus, device, and storage medium for evaluating a crowded merging scenario, which are used to improve the accuracy of evaluating how autonomous driving software handles crowded merging scenarios, improve the efficiency of autonomous driving testing, and improve the applicability and safety of autonomous driving.

[0029] 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.

[0030] 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 congested merging scenario in an embodiment of the present invention includes:

[0031] 101. Obtain a vehicle driving scene dataset corresponding to a target vehicle. The vehicle driving scene dataset is used to indicate scene data collected during the target vehicle's driving on an actual road scene.

[0032] The target vehicle has an autonomous driving function, and the vehicle driving scene dataset is used to indicate the scene data collected by the target vehicle during driving in actual road scenes. That is, while the target vehicle is driving in actual road scenes, the positioning module, perception module, and data acquisition module are activated to record the target vehicle's location data, speed data, acceleration data, and obstacle vehicles and other obstacles in real time, thereby obtaining the vehicle driving scene dataset. At the same time, in order to increase the probability of a congested merging scenario, the target vehicle needs to travel in a road scene with fewer lane lines and / or at a congested road intersection to achieve a merging scenario. For example, the target vehicle travels in the rightmost lane of the through lane and merges with an obstacle vehicle turning right, or the target vehicle travels in the right-turn lane and merges with a vehicle in the through lane.

[0033] Specifically, the server receives a data upload request sent by the target vehicle, which includes a vehicle driving scene dataset; the server reads the vehicle driving scene dataset from the data upload request; the server updates the vehicle driving scene dataset to a preset distributed database cluster through a read-write separation method, thereby improving data reading and writing efficiency.

[0034] It is understandable that the execution subject of the present invention may be a congested merging 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.

[0035] 102. Perform data cleaning and data framing processing on the vehicle driving scene data set in sequence using a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data.

[0036] It is understood that the preset cleaning strategy and preset framing rules are both pre-stored in a preset configuration data table. The preset cleaning strategy is used to indicate the cleaning standards and data cleaning specifications for the vehicle driving scene dataset, and the preset framing rules are used to indicate the method for splitting the vehicle driving scene dataset. In some real-time methods, the server reads the preset cleaning strategy and preset framing rules from the preset configuration data table; the server uses the preset cleaning strategy to sequentially fill missing data, correct erroneous data, remove data duplication, unify the data format, and cluster the data on the vehicle driving scene dataset in the preset distributed database cluster, thereby obtaining cleaned vehicle driving scene data; the server then performs non-overlapping framing on the cleaned vehicle driving scene data according to the preset framing rules, thereby obtaining multiple framed scene data. In other words, the time periods corresponding to the respective framed scene data do not overlap, and the vehicle driving scenes corresponding to the respective framed scene data differ from each other.

[0037] In some embodiments, the server performs scene annotation processing on the cleaned vehicle driving scene data using a preset scene annotation model to obtain annotated vehicle driving scene data. The server then performs data framing processing on the annotated vehicle driving scene data according to preset framing rules to obtain multiple framed scene data. The preset scene annotation model is used to indicate a pre-trained scene semantic annotation model. The preset scene annotation model can be a supervised learning model based on deep learning or a convolutional neural network model, and the specifics are not limited here.

[0038] 103. Perform lane merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested lane merging scene data.

[0039] It is understandable that the number of congested merging scenarios corresponding to the target congested merging scenario data may be one or more, and the specific number is not limited here. Figure 2 As shown, lanes B and C merge into lane A at intersection P, and lane D is an adjacent lane to lanes A and B. Therefore, vehicles traveling in lanes B and C can merge. Furthermore, when any of lanes A, B, or C is a busy lane, vehicles traveling in lanes B and C can merge in a congested manner. Lanes A and B are topologically separated by intersection P and are therefore marked as two different lanes in the diagram. Vehicles in lanes A and B travel in essentially the same direction and lane.

[0040] Specifically, the server determines between any two frames of the lane where the reference obstacle vehicle is located in each framed scene data within a preset time period according to a preset scene screening algorithm. If there is a difference between any two frames of the lane where the reference obstacle vehicle is located in each framed scene data within the preset time period, the server determines whether the reference obstacle vehicle in each framed scene data travels from the first lane to the merging lane (that is, the lane where the target position is located), and the first lane is the lane where the reference obstacle vehicle travels before reaching the merging lane, and the first lane is not an adjacent lane to the merging lane, for example, Figure 2 As shown, the merging lane is lane A, the first lane can be lane B or C, and the first lane is not lane D; if the reference obstacle vehicle in each framed scene data is traveling from the first lane to the merging lane, the server determines whether the target vehicle in each framed scene data is traveling from the second lane to the merging lane, the second lane being the lane the target vehicle is traveling before reaching the merging lane, and the second lane is not an adjacent lane to the merging lane, for example, Figure 2 As shown, the merging lane is lane A, the second lane can be lane B or lane C, and the second lane is not lane D; if the target vehicle in each framed scene data is traveling from the second lane to the merging lane, the server determines whether the first lane and the second lane of each framed scene data are different lanes. For example, if the first lane is lane B and the second lane is lane C, the server determines that the first lane and the second lane are different lanes; if the first lane and the second lane in each framed scene data are different lanes, the server determines whether the time difference between the target vehicle and the reference obstacle vehicle in each framed scene data when they arrive at the merging lane is less than a preset time threshold; if the time difference between the target vehicle and the reference obstacle vehicle when they arrive at the merging lane is less than the preset time threshold, the server determines that the target vehicle whether the target distance difference between the target vehicle and the reference obstacle vehicle reaching the merging lane in each framed scene data is less than the preset distance difference; if the target distance difference between the target vehicle and the reference obstacle vehicle reaching the merging lane in each framed scene data is less than the preset distance difference, the server determines whether the number of other obstacle vehicles within the preset distance in the first lane or the second lane in each framed scene data is greater than the preset number threshold; if the number of other obstacle vehicles within the preset distance in the first lane or the second lane in each framed scene data is greater than the preset number threshold, the server obtains the corresponding congested merging scene data from each framed scene data according to the preset data extraction time period, and merges the congested merging scene data corresponding to each framed scene data according to the overlapping time period to obtain the target congested merging scene data.

[0041] It should be noted that the server pushes the moment when the target vehicle and the reference obstacle vehicle are closest to each other forward by a short period of time (for example, 10 seconds) as the starting extraction moment, and takes the moment when the target vehicle and the reference obstacle vehicle last arrive at the merge lane as the ending extraction moment; the server sets the time period between the ending extraction moment and the starting extraction moment as the preset data extraction time period.

[0042] 104. Perform scenario simulation and autonomous driving behavior evaluation based on the target congested lane merging scenario data to obtain evaluation results.

[0043] Specifically, the server starts the preset autonomous driving software in a preset simulation environment and simulates the target congested lane merging scenario data through the autonomous driving software. When the preset autonomous driving software receives the positioning data and perception information in the target congested lane merging scenario data, it generates a preset control instruction and controls the autonomous driving vehicle to perform a congested lane merging according to the preset control instruction, thereby obtaining the corresponding congested lane merging simulation behavior data of the autonomous driving vehicle. The server extracts the corresponding actual congested lane merging behavior data of the target vehicle from the target congested lane merging scenario data. The server compares whether there is a difference between the congested lane merging simulation behavior data and the actual congested lane merging behavior data, obtains the difference data, and determines an evaluation result based on the difference data. The evaluation result is used to indicate the safety level of the preset autonomous driving software. The safety level may include level 0, level 1, level 2, level 3, etc., which are not specifically limited here. The higher the safety level value, the better the safety performance of the preset autonomous driving software.

[0044] In some embodiments, a server extracts the start time and completion time of the crowded merge simulation from the crowded merge simulation behavior data, and reads the actual start time and completion time of the crowded merge from the actual crowded merge behavior data. The server obtains a first difference value between the start time and the actual start time of the crowded merge, and a second difference value between the completion time and the actual start time of the crowded merge. The server searches a preset scoring rule based on the first difference value and the second difference value to obtain a score value corresponding to each of the first difference value and the second difference value. The server calculates the sum of the score values ​​corresponding to each of the first difference value and the second difference value, and retrieves a preset security level configuration data table based on the sum of the score values ​​to obtain an evaluation result. The first difference value and the second difference value are both difference data. For example, when the first difference value and the second difference value are both 0, the server determines that the sum of the scores corresponding to the first difference value and the second difference value is 2, and the server sets the evaluation result to a security level of 5. When the first difference value is 0.4 and the second difference value is 0.6, the server determines that the sum of the scores corresponding to the first difference value and the second difference value is 1, and the server sets the evaluation result to a security level of 2.

[0045] It should be noted that the greater the difference between the actual congested merging behavior data and the actual congested merging behavior data, the lower the corresponding safety level of the evaluation result. When the actual congested merging behavior data and the actual congested merging behavior data are exactly the same, the score is full, and the corresponding safety level of the evaluation result is the highest. If the target congested merging scenario data includes multiple scenarios, the server also needs to compare the actual congested merging behavior data and the actual congested merging behavior data in each scenario to obtain the difference data for each scenario, and calculate the corresponding score value based on the difference data for each scenario. The server then evaluates the preset autonomous driving software's ability to handle congested merging scenarios based on the corresponding score value for each scenario, and obtains the evaluation result.

[0046] In an embodiment of the present invention, target congested merging scene data is extracted from a vehicle driving scene data set through a preset cleaning strategy, preset framing rules, and a preset scene screening algorithm, thereby improving the accuracy of screening the target congested merging scene data. Scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain evaluation results, thereby improving the accuracy of evaluating the autonomous driving software's handling of congested merging scenarios, improving the efficiency of autonomous driving testing, and improving the applicability and safety of autonomous driving.

[0047] See also Figure 3 Another embodiment of the method for evaluating a congested merging scenario in the embodiment of the present invention includes:

[0048] 301. Obtain a vehicle driving scene dataset corresponding to a target vehicle, where the vehicle driving scene dataset is used to indicate scene data collected during the target vehicle's driving on an actual road scene.

[0049] The specific execution process of step 301 is similar to the specific execution process of step 101, and will not be repeated here.

[0050] 302. Perform data cleaning and data framing processing on the vehicle driving scene data set in sequence using a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data.

[0051] The specific execution process of step 302 is similar to the specific execution process of step 102, and will not be repeated here.

[0052] 303. Perform vehicle driving trajectory analysis on each framed scene data according to a preset scene screening algorithm to obtain a lane position relationship corresponding to each framed scene data. The lane position relationship is used to indicate relative vehicle lane position information between the target vehicle and the reference obstacle vehicle.

[0053] Among them, the lane position relationship is used to indicate the relative vehicle lane position information between the target vehicle and the reference obstacle vehicle, such as Figure 2As shown, for example, in the same framed scene data, the target vehicle is traveling in lane D and the reference obstacle vehicle is traveling in lane B. The server determines that there is no lane merging behavior between the target vehicle and the reference obstacle vehicle, and the lane position relationship corresponding to the framed scene data is a null value.

[0054] Specifically, the server extracts the reference obstacle vehicle corresponding to the target vehicle from each framed scene data according to a preset scene screening algorithm, and traverses each frame of data for the reference obstacle vehicle within a preset time range. If the lane where the reference obstacle vehicle is located in the current frame is different from the lane where it was located in the previous frame, the server determines whether the lane where the reference obstacle vehicle is located in the current frame is a merging lane. The merging lane is a lane formed by the merging of two or more lanes, such as Figure 2 As shown, lane A is the merging lane; if the lane where the reference obstacle vehicle is currently located is the merging lane, the server determines whether the reference obstacle vehicle has traveled from the first lane to the merging lane (that is, the lane where the target position is located). The first lane is the lane that the reference obstacle vehicle traveled before reaching the merging lane, and the first lane is not an adjacent lane to the merging lane. Figure 2 As shown, for example, the first lane is lane B or lane C; if the reference obstacle vehicle is traveling from the first lane to the merging lane, then it is determined whether the target vehicle is traveling from the second lane to the merging lane, and the second lane is the lane the target vehicle was traveling before reaching the merging lane, and the second lane is not an adjacent lane to the merging lane, such as Figure 2 As shown, for example, the second lane is lane B or lane C; if the target vehicle is traveling from the second lane to the merging lane, the server determines whether the first lane and the second lane are different lanes, such as Figure 2 As shown, for example, the first lane and the second lane are lane B or lane C respectively; if the first lane and the second lane are different lanes, the server determines that each framed scene data includes a congested merging scene, and the server generates relative vehicle lane position information between the target vehicle and the reference obstacle vehicle, and sets the relative vehicle lane position information between the target vehicle and the reference obstacle vehicle as the lane position relationship corresponding to each framed scene data.

[0055] 304. Determine whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle according to the lane position relationship corresponding to each framed scene data. The target lane merging behavior includes left merging behavior and right merging behavior.

[0056] The target merging behavior includes left merging behavior and right merging behavior. In some embodiments, the server determines whether the target vehicle and the reference obstacle vehicle meet preset merging scenario correlation rules based on the lane position relationship corresponding to each framed scene data. The preset merging scenario correlation rules are used to indicate the relevant criteria for the mutual influence between the target vehicle and the reference obstacle vehicle when the merging behavior occurs. That is, in the merging scenarios selected by the server using the preset merging scenario correlation rules, the target vehicle and the reference obstacle vehicle are close to each other and there is an influence between the two vehicles, so that the target congested merging scenario data is more meaningful. If the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rules, the server determines whether the lane where the reference obstacle vehicle was located before merging was a congested lane. If the lane where the reference obstacle vehicle was located before merging was a congested lane, the server determines that a target merging behavior occurred between the target vehicle and the reference obstacle vehicle.

[0057] Furthermore, when the server is executing the step of determining whether the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rule according to the lane position relationship corresponding to each framed scene data, in some embodiments, the server obtains the first moment corresponding to when the target vehicle travels to the target position, the second moment corresponding to when the reference obstacle vehicle travels to the target position, and the target distance difference according to the lane position relationship corresponding to each framed scene data. The target distance difference is used to indicate the minimum distance between the target vehicle and the reference obstacle vehicle before they reach the target position; the server calculates the target moment difference based on the first moment and the second moment; when the target moment difference is less than the preset time threshold, and the target distance difference is less than the preset distance difference, the server determines that the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rule. For example, Figure 2 As shown, the target vehicle and the reference obstacle vehicle are traveling in lane B and lane C, respectively, before merging. If the target time difference between the target vehicle and the reference obstacle vehicle at the target position A (i.e., the merging lanes of the target vehicle and the reference obstacle vehicle) is less than a preset time threshold (e.g., the preset time threshold is 5 seconds), and the target distance difference between the target vehicle and the reference obstacle vehicle before reaching the target position A is less than a preset distance difference (e.g., the preset distance difference is 5 meters), the server determines that the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule. If the target time difference between the target vehicle and the reference obstacle vehicle at the target position A is greater than or equal to the preset time threshold, and / or the target distance difference between the target vehicle and the reference obstacle vehicle before reaching the target position A is greater than or equal to the preset distance difference, the server determines that the target vehicle and the reference obstacle vehicle do not meet the preset merging scenario correlation rule.

[0058] Furthermore, when the server performs the step of determining whether the lane in which the reference obstacle vehicle was merging before merging is a congested lane if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, in some embodiments, if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, the server obtains the number of other obstacle vehicles within a preset distance of the target lane, which is the lane in which the reference obstacle vehicle was merging before merging. The server then determines whether the number of other obstacle vehicles is greater than a preset number threshold. If the number of other obstacle vehicles is greater than the preset number threshold, the server determines that the lane in which the reference obstacle vehicle was merging before merging is a congested lane. For example, if the number of other obstacle vehicles within a preset distance (e.g., 50 meters) of the lane in which the reference obstacle vehicle was merging before merging (e.g., lane B or lane C) is greater than a preset number threshold (e.g., the preset number threshold is 4), the server determines that the lane in which the reference obstacle vehicle was merging before merging is a congested lane. The server further determines that a target merging behavior exists between the target vehicle and the reference obstacle vehicle, and the server executes step 305.

[0059] 305. If there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle, generate target congested lane merging scene data based on each framed scene data corresponding to the target lane merging behavior.

[0060] In some embodiments, if there is a target merging behavior between the target vehicle and the reference obstacle vehicle, the server sets a congested merging label for each framed scene data corresponding to the target merging behavior; when the server has completed marking multiple framed scene data, the server filters the multiple framed scene data according to the congested merging label to obtain filtered scene data, and combines the filtered scene data into target congested merging scene data. The server sets a congested merging scene identifier, and stores the target congested merging scene data in a preset scene data table according to the congested merging scene identifier.

[0061] 306. Perform scenario simulation and autonomous driving behavior evaluation based on the target congested lane merging scenario data to obtain an evaluation result.

[0062] That is, after simulating a crowded merging scenario based on the target crowded merging scenario data, the server evaluates and processes the autonomous driving behavior in the simulated crowded merging scenario to obtain an evaluation result. The evaluation result is used to indicate the safety level of the preset autonomous driving software. In some embodiments, the server simulates the target crowded merging scenario data using preset autonomous driving simulation software to obtain autonomous driving crowded merging simulation scenario data; the server triggers the autonomous driving vehicle in the autonomous driving crowded merging simulation scenario data according to preset control instructions to obtain the corresponding crowded merging simulation behavior data of the autonomous driving vehicle; the server extracts the corresponding actual crowded merging behavior data of the target vehicle from the target crowded merging scenario data, and performs scenario evaluation processing on the simulated crowded merging behavior data and the actual crowded merging behavior data according to preset scenario scoring rules to obtain an evaluation result. Furthermore, the server calculates a target displacement difference value between the actual driving behavior and the automatic driving behavior based on the crowded merging simulation behavior data and the actual crowded merging behavior data, wherein the target displacement difference value between the actual driving behavior and the automatic driving behavior includes a first displacement difference value projected along the lane line direction and a second displacement difference value projected perpendicular to the lane line direction; the server calculates a crowded merging score corresponding to the first displacement difference value and a crowded merging score corresponding to the second displacement difference value according to a preset scenario scoring rule; the server calculates a total score of the crowded merging simulation behavior based on the crowded merging score corresponding to the first displacement difference value and the crowded merging score corresponding to the second displacement difference value, and queries a preset safety level configuration data table based on the total score of the crowded merging simulation behavior to obtain an evaluation result.

[0063] In some embodiments, after step 306, the server generates a congested merging scenario evaluation report based on the preset autonomous driving simulation software, the target congested merging scenario data, and the evaluation results; the server pushes the congested merging scenario evaluation report to the target terminal, so that the target terminal displays the congested merging scenario evaluation report in a visual chart. Specifically, the server obtains a congested merging scenario template, and the server performs data conversion and report generation processing on the preset autonomous driving simulation software, the target congested merging scenario data, and the evaluation results to obtain the congested merging scenario evaluation report, and stores the congested merging scenario evaluation report in a preset file system; the server receives a congested merging scenario evaluation report display request, and obtains a file identifier from the congested merging scenario evaluation report request. The server obtains the congested merging scenario evaluation report from the preset file system according to the file identifier, and calls a preset file push interface to push the congested merging scenario evaluation report to the target terminal, so that the target terminal displays the congested merging scenario evaluation report in a visual chart.

[0064] It is understood that the server's evaluation results can objectively and accurately assess the preset autonomous driving simulation software's handling of congested merging scenarios and be used to optimize the preset autonomous driving simulation software's output. Furthermore, when comparing different versions of autonomous driving software, the server uses the same data (i.e., the same target congested merging scenario data) for software simulation and autonomous driving behavior evaluation, thereby eliminating the problem of significant deviations in evaluation results between simulations using different test data sets.

[0065] In an embodiment of the present invention, target congested merging scene data is extracted from a vehicle driving scene data set through a preset cleaning strategy, preset framing rules, and a preset scene screening algorithm, thereby improving the accuracy of screening the target congested merging scene data. Scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain evaluation results, thereby improving the accuracy of evaluating the autonomous driving software's handling of congested merging scenarios, improving the efficiency of autonomous driving testing, and improving the applicability and safety of autonomous driving.

[0066] The above describes the method for evaluating a crowded merging scene in an embodiment of the present invention. The following describes the device for evaluating a crowded merging scene in an embodiment of the present invention. Figure 4 In one embodiment of the present invention, a device for evaluating a congested merging scenario includes:

[0067] An acquisition module 401 is configured to acquire a vehicle driving scene dataset corresponding to a target vehicle, wherein the vehicle driving scene dataset is configured to indicate scene data collected during the target vehicle's driving on an actual road scene;

[0068] The processing module 402 is used to perform data cleaning and data framing processing on the vehicle driving scene dataset in sequence according to a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data;

[0069] A screening module 403 is configured to perform merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested merging scene data;

[0070] The evaluation module 404 is used to perform scenario simulation and autonomous driving behavior evaluation processing based on the target congested lane merging scenario data to obtain an evaluation result.

[0071] In an embodiment of the present invention, target congested merging scene data is extracted from a vehicle driving scene data set through a preset cleaning strategy, preset framing rules, and a preset scene screening algorithm, thereby improving the accuracy of screening the target congested merging scene data. Scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain evaluation results, thereby improving the accuracy of evaluating the autonomous driving software's handling of congested merging scenarios, improving the efficiency of autonomous driving testing, and improving the applicability and safety of autonomous driving.

[0072] See also Figure 5 Another embodiment of the congested merging scenario assessment device according to the embodiment of the present invention includes:

[0073] An acquisition module 401 is configured to acquire a vehicle driving scene dataset corresponding to a target vehicle, wherein the vehicle driving scene dataset is configured to indicate scene data collected during the target vehicle's driving on an actual road scene;

[0074] The processing module 402 is used to perform data cleaning and data framing processing on the vehicle driving scene dataset in sequence according to a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data;

[0075] A screening module 403 is configured to perform merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested merging scene data;

[0076] The evaluation module 404 is used to perform scenario simulation and autonomous driving behavior evaluation processing based on the target congested lane merging scenario data to obtain an evaluation result.

[0077] In a feasible implementation manner, the screening module 403 further includes:

[0078] An analysis unit 4031 is configured to perform vehicle driving trajectory analysis on each framed scene data according to a preset scene screening algorithm to obtain a lane position relationship corresponding to each framed scene data, wherein the lane position relationship indicates relative vehicle lane position information between the target vehicle and a reference obstacle vehicle;

[0079] A judgment unit 4032 is configured to judge whether a target lane merging behavior exists between the target vehicle and the reference obstacle vehicle based on the lane position relationship corresponding to each framed scene data, wherein the target lane merging behavior includes a left merging behavior and a right merging behavior;

[0080] The generating unit 4033 is configured to generate target congested lane merging scene data based on each framed scene data corresponding to the target lane merging behavior if there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle.

[0081] In a feasible implementation manner, the determining unit 4032 further includes:

[0082] The first judgment subunit 40321 is configured to judge whether the target vehicle and the reference obstacle vehicle meet a preset merging scenario correlation rule based on the lane position relationship corresponding to each framed scene data;

[0083] The second judgment subunit 40322 is configured to determine whether the lane in which the reference obstacle vehicle was located before merging is a congested lane if the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule;

[0084] The determining subunit 40323 is configured to determine that a target merging behavior exists between the target vehicle and the reference obstacle vehicle if the lane in which the reference obstacle vehicle was located before merging is a congested lane.

[0085] In a feasible implementation manner, the first judgment subunit 40321 is specifically configured to:

[0086] According to the lane position relationship corresponding to each framed scene data, obtaining a first time corresponding to when the target vehicle travels to the target position, a second time corresponding to when the reference obstacle vehicle travels to the target position, and a target distance difference, wherein the target distance difference indicates the minimum distance between the target vehicle and the reference obstacle vehicle before reaching the target position;

[0087] Calculating a target time difference according to the first time and the second time;

[0088] When the target time difference is less than a preset time threshold, and the target distance difference is less than a preset distance difference, it is determined that the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rule.

[0089] In a feasible implementation manner, the second judgment subunit 40322 is specifically configured to:

[0090] If the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, the number of other obstacle vehicles within a preset distance in the target lane is obtained, where the target lane is the lane where the reference obstacle vehicle was before merging.

[0091] Determining whether the number of vehicles with other obstacles is greater than a preset number threshold;

[0092] If the number of the other obstacle vehicles is greater than a preset number threshold, it is determined that the lane where the reference obstacle vehicle was located before merging is a congested lane.

[0093] In a feasible implementation, the evaluation module 404 is specifically configured to:

[0094] Performing a lane merging scenario simulation on the target congested lane merging scenario data using preset autonomous driving simulation software to obtain autonomous driving congested lane merging simulation scenario data;

[0095] triggering the autonomous driving vehicle in the autonomous driving congested lane merging simulation scene data according to a preset control instruction to obtain congested lane merging simulation behavior data corresponding to the autonomous driving vehicle;

[0096] The actual crowded merging behavior data corresponding to the target vehicle is extracted from the target crowded merging scene data, and a scene evaluation process is performed on the simulated crowded merging behavior data and the actual crowded merging behavior data according to a preset scene scoring rule to obtain an evaluation result.

[0097] In a feasible implementation manner, the congested merging scenario assessment device further includes:

[0098] A generating module 405 is configured to generate a congested merging scenario assessment report based on the preset autonomous driving simulation software, the target congested merging scenario data, and the assessment result;

[0099] The push module 406 is configured to push the congested merging scenario assessment report to a target terminal, so that the target terminal performs a visual chart display according to the congested merging scenario assessment report.

[0100] In an embodiment of the present invention, target congested merging scene data is extracted from a vehicle driving scene data set through a preset cleaning strategy, preset framing rules, and a preset scene screening algorithm, thereby improving the accuracy of screening the target congested merging scene data. Scene simulation and autonomous driving behavior evaluation processing are performed based on the target congested merging scene data to obtain evaluation results, thereby improving the accuracy of evaluating the autonomous driving software's handling of congested merging scenarios, improving the efficiency of autonomous driving testing, and improving the applicability and safety of autonomous driving.

[0101] above Figure 4 and Figure 5 The congested merging scene assessment apparatus in the embodiment of the present invention is described in detail from a modular perspective. The congested merging scene assessment device in the embodiment of the present invention is described in detail from a hardware processing perspective.

[0102] Figure 6: is a structural diagram of a crowded merging scene assessment device provided by an embodiment of the present invention. The crowded merging scene assessment device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be temporary storage or permanent storage. The program stored in the storage medium 630 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 merging scene assessment device 600. Furthermore, the processor 610 can be configured to communicate with the storage medium 630 to execute a series of computer program operations in the storage medium 630 on the crowded merging scene assessment device 600.

[0103] The congested merging scenario assessment device 600 may further include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 6 The structure of the crowded merging scene assessment device shown does not constitute a limitation on the crowded merging scene assessment device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0104] 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 method for evaluating a crowded merging scenario.

[0105] 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.

[0106] 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.

[0107] 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 merging scenario, characterized in that: The congested merging scenario assessment method includes: Acquire a vehicle driving scene data set corresponding to a target vehicle, wherein the vehicle driving scene data set is used to indicate scene data collected during the target vehicle's driving on an actual road scene; Performing data cleaning and data framing processing on the vehicle driving scene data set in sequence according to a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; Performing merging scene recognition and vehicle congestion detection on the plurality of framed scene data respectively according to a preset scene screening algorithm to obtain target congested merging scene data; Performing scenario simulation and autonomous driving behavior evaluation processing based on the target congested merging scenario data to obtain an evaluation result; The method of performing scenario simulation and autonomous driving behavior evaluation based on the target congested merging scenario data to obtain an evaluation result includes: controlling the autonomous driving vehicle to perform a congested merging according to a preset control instruction to obtain congested merging simulation behavior data corresponding to the autonomous driving vehicle; extracting actual congested merging behavior data corresponding to the target vehicle from the target congested merging scenario data; extracting a congested merging simulation start time and a congested merging simulation completion time from the congested merging simulation behavior data; and reading an actual congested merging start time and an actual congested merging end time from the actual congested merging behavior data; respectively obtaining a first difference value between the congested merging simulation start time and the actual congested merging start time, and a second difference value between the congested merging simulation completion time and the actual congested merging start time; searching a preset scoring rule based on the first difference value and the second difference value to obtain score values ​​corresponding to each of the first difference value and the second difference value; calculating a sum of the score values ​​corresponding to each of the first difference value and the second difference value; and retrieving a preset safety level configuration data table based on the sum of the score values ​​to obtain an evaluation result.

2. The method for evaluating a crowded merging scenario according to claim 1, characterized in that: The performing lane merging scene recognition and vehicle congestion detection on the plurality of framed scene data according to a preset scene screening algorithm to obtain target congested lane merging scene data includes: Performing vehicle driving trajectory analysis on each framed scene data according to a preset scene screening algorithm to obtain a lane position relationship corresponding to each framed scene data, wherein the lane position relationship is used to indicate relative vehicle lane position information between the target vehicle and the reference obstacle vehicle; Determining whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle according to the lane position relationship corresponding to each framed scene data, wherein the target lane merging behavior includes a left lane merging behavior and a right lane merging behavior; If there is a target merging behavior between the target vehicle and the reference obstacle vehicle, target congested merging scene data is generated based on each framed scene data corresponding to the target merging behavior.

3. The method for evaluating a crowded merging scenario according to claim 2, wherein: The determining, based on the lane position relationship corresponding to each framed scene data, whether there is a target lane merging behavior between the target vehicle and the reference obstacle vehicle includes: According to the lane position relationship corresponding to each framed scene data, determining whether the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rules; If the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, determining whether the lane where the reference obstacle vehicle was located before merging is a congested lane; If the lane where the reference obstacle vehicle was located before merging is a congested lane, it is determined that there is a target merging behavior between the target vehicle and the reference obstacle vehicle.

4. The method for evaluating a crowded merging scenario according to claim 3, wherein: The determining, based on the lane position relationship corresponding to each framed scene data, whether the target vehicle and the reference obstacle vehicle meet a preset merging scene correlation rule includes: According to the lane position relationship corresponding to each framed scene data, obtaining a first time corresponding to when the target vehicle travels to the target position, a second time corresponding to when the reference obstacle vehicle travels to the target position, and a target distance difference, wherein the target distance difference indicates the minimum distance between the target vehicle and the reference obstacle vehicle before reaching the target position; Calculating a target time difference according to the first time and the second time; When the target time difference is less than a preset time threshold, and the target distance difference is less than a preset distance difference, it is determined that the target vehicle and the reference obstacle vehicle meet the preset merging scene correlation rule.

5. The method for evaluating a crowded merging scenario according to claim 3, wherein: If the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, determining whether the lane where the reference obstacle vehicle was located before merging is a congested lane includes: If the target vehicle and the reference obstacle vehicle meet the preset merging scenario correlation rule, the number of other obstacle vehicles within a preset distance in the target lane is obtained, where the target lane is the lane where the reference obstacle vehicle was before merging. Determining whether the number of vehicles with other obstacles is greater than a preset number threshold; If the number of the other obstacle vehicles is greater than a preset number threshold, it is determined that the lane where the reference obstacle vehicle was located before merging is a congested lane.

6. The method for evaluating a congested merging scenario according to any one of claims 1 to 5, wherein: After performing scenario simulation and autonomous driving behavior evaluation processing based on the target congested merging scenario data to obtain an evaluation result, the congested merging scenario evaluation method further includes: generating a congested merging scenario assessment report based on the preset autonomous driving simulation software, the target congested merging scenario data, and the assessment result; The congested merging scenario assessment report is pushed to a target terminal, so that the target terminal displays a visual chart according to the congested merging scenario assessment report.

7. A device for evaluating a crowded merging scene, characterized in that: The congested merging scene assessment device comprises: An acquisition module is used to acquire a vehicle driving scene data set corresponding to a target vehicle, wherein the vehicle driving scene data set is used to indicate scene data collected during the target vehicle's driving on an actual road scene; a processing module, configured to sequentially perform data cleaning and data framing processing on the vehicle driving scene dataset using a preset cleaning strategy and preset framing rules to obtain a plurality of framed scene data; a screening module, configured to perform merging scene recognition and vehicle congestion detection on the plurality of framed scene data respectively according to a preset scene screening algorithm, to obtain target congested merging scene data; an evaluation module, configured to perform scenario simulation and autonomous driving behavior evaluation processing based on the target congested merging scenario data to obtain an evaluation result; The method of performing scenario simulation and autonomous driving behavior evaluation based on the target congested merging scenario data to obtain an evaluation result includes: controlling the autonomous driving vehicle to perform a congested merging according to a preset control instruction to obtain congested merging simulation behavior data corresponding to the autonomous driving vehicle; extracting actual congested merging behavior data corresponding to the target vehicle from the target congested merging scenario data; extracting a congested merging simulation start time and a congested merging simulation completion time from the congested merging simulation behavior data; and reading an actual congested merging start time and an actual congested merging end time from the actual congested merging behavior data; respectively obtaining a first difference value between the congested merging simulation start time and the actual congested merging start time, and a second difference value between the congested merging simulation completion time and the actual congested merging start time; searching a preset scoring rule based on the first difference value and the second difference value to obtain score values ​​corresponding to each of the first difference value and the second difference value; calculating a sum of the score values ​​corresponding to each of the first difference value and the second difference value; and retrieving a preset safety level configuration data table based on the sum of the score values ​​to obtain an evaluation result.

8. A device for evaluating a crowded merging scenario, characterized in that: The congested merging scenario assessment device comprises: 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 congested merging scene assessment device to execute the congested merging scene assessment method according to any one of claims 1 to 6.

9. 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 merging scene according to any one of claims 1 to 6 is implemented.

Images