CN-121980211-A - Performance evaluation method and device for road side algorithm, electronic equipment and storage medium

CN121980211ACN 121980211 ACN121980211 ACN 121980211ACN-121980211-A

Abstract

The application discloses a performance evaluation method and device of a road side algorithm, electronic equipment and a storage medium, and belongs to the technical field of vehicle-road cooperation. The method comprises the steps of obtaining a track sequence output by a road side algorithm, wherein each track point in the track sequence comprises position information and kinematic features, carrying out mutation detection on the kinematic features of adjacent track points in the track sequence to obtain a mutation detection result, carrying out anomaly detection on each track point in the track sequence according to the position information and the kinematic features to obtain an anomaly detection result, and determining a performance evaluation result of the road side algorithm according to the mutation detection result and the anomaly detection result. The method and the device can comprehensively evaluate the performance of the road side algorithm, avoid the limitation of single index evaluation, and comprehensively reflect the comprehensive performance of the road side algorithm in a complex traffic environment.

Inventors

LI BOXIAN
ZHANG JIE
XU WENLONG
FENG WENQIAN
LEI YANG

Assignees

中信科智联科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251202

Claims (15)

1. A performance evaluation method for a roadside algorithm, comprising: Acquiring a track sequence output by a road side algorithm, wherein each track point in the track sequence comprises position information and kinematic characteristics; Mutation detection is carried out on the kinematic characteristics of adjacent track points in the track sequence, so that a mutation detection result is obtained; according to the position information and the kinematic characteristics, carrying out anomaly detection on each track point in the track sequence to obtain an anomaly detection result; And determining a performance evaluation result of the road side algorithm according to the mutation detection result and the abnormality detection result.
2. The method according to claim 1, wherein the mutation detection of the kinematic features of adjacent track points in the track sequence to obtain a mutation detection result includes: determining a change value of the kinematic features of the adjacent track points; If the absolute value of the change value is larger than or equal to a change threshold value, determining that a kinematic mutation exists in a next track point in the adjacent track points, and recording the mutation type and the change value of the next track point to obtain the mutation detection result.
3. The method of claim 2, wherein the kinematic features include at least one of speed, acceleration, and heading angle, and the abrupt type includes at least one of speed abrupt change, acceleration abrupt change, and heading angle abrupt change.
4. A method according to claim 3, further comprising at least one of: generating a first graph of the speed over time; Generating a second graph of the acceleration over time; And before the third curve graph is generated, if the initial course angle of the track point is larger than 180 degrees relative to the course angle change value of the previous track point, subtracting 360 degrees from the initial course angle of the track point to be used as the final course angle of the track point, and if the initial course angle of the track point is smaller than-180 degrees relative to the course angle change value of the previous track point, adding 360 degrees to the initial course angle of the track point to be used as the final course angle of the track point.
5. The method according to claim 1, wherein the performing anomaly detection on each track point in the track sequence according to the position information and the kinematic feature to obtain an anomaly detection result includes: determining a burr point in the track sequence according to the position information and the kinematic characteristics; determining a breakpoint in the track sequence according to the position information and the speed in the kinematic feature; determining smoothness abnormal points in the track sequence based on a sliding window; And determining the abnormality detection result according to the burr point, the breakpoint and the smoothness abnormality point.
6. The method of claim 5, wherein the determining the burr point in the sequence of trajectories based on the location information and the kinematic feature comprises at least one of: Determining the curvature of each track point according to the position information, and determining the track point with the curvature meeting the target condition as the burr point; Determining track points with the change quantity of the course angle in the kinematic features larger than a course angle change threshold value as the burr points; clustering each track point in the track sequence based on a DBSCAN clustering algorithm, and determining the isolated track point in the clustering result as the burr point.
7. The method of claim 6, wherein the target condition comprises a curvature of the locus point being greater than or equal to a curvature threshold and a curvature of both a preceding locus point and a following locus point of the locus point being greater than or equal to a target proportion of the curvature threshold.
8. The method of claim 5, wherein determining a breakpoint in the sequence of trajectories based on the location information and the velocity in the kinematic feature comprises: and determining the track point with the speed ratio between the speed and the speed of the previous track point being greater than or equal to a speed ratio threshold value and the distance between the track point and the previous track point being greater than or equal to a distance threshold value as the breakpoint.
9. The method of claim 5, wherein determining the anomaly detection result based on the burr point, the breakpoint, and the smoothness anomaly point comprises: And filtering repeated track points in the burr points, the break points and the smoothness abnormal points, and determining the abnormal detection result according to the filtered burr points, break points and smoothness abnormal points.
10. The method of claim 9, wherein determining the anomaly detection result based on the filtered burr points, break points, and smoothness anomaly points comprises: Counting the number of the burr points, the number of the break points, the burr rate and the breakpoint rate from the filtered burr points, the break points and the smoothness abnormal points, wherein the burr rate is the occupied ratio of the burr points in the filtered burr points, the break points and the smoothness abnormal points in all track points, and the breakpoint rate is the occupied ratio of the break points in the filtered burr points, the break points and the smoothness abnormal points in the track points; And taking the number of the burr points, the number of the break points, the burr rate and the break point rate as the abnormality detection result.
11. The method of claim 5, further comprising, after said determining said anomaly detection result from said burr point, said breakpoint, and said smoothness anomaly point: and generating a visualized result of the track sequence according to the abnormality detection result.
12. The method according to any one of claims 1-11, further comprising: acquiring a reference course of a lane where a first track point in the track sequence is located; determining whether the course angle of the first track point is accurate according to the reference course to obtain an initial course angle detection result; the determining the performance evaluation result of the road side algorithm according to the mutation detection result and the abnormality detection result includes: And determining a performance evaluation result of the road side algorithm according to the mutation detection result, the abnormality detection result and the initial course angle detection result.
13. A performance evaluation device for a roadside algorithm, comprising: The track acquisition module is used for acquiring a track sequence output by the road side algorithm, and each track point in the track sequence comprises position information and kinematic characteristics; the mutation detection module is used for carrying out mutation detection on the kinematic characteristics of adjacent track points in the track sequence to obtain a mutation detection result; the anomaly detection module is used for carrying out anomaly detection on each track point in the track sequence according to the position information and the kinematic characteristics to obtain an anomaly detection result; And the evaluation result determining module is used for determining the performance evaluation result of the road side algorithm according to the mutation detection result and the abnormality detection result.
14. An electronic device comprising a processor, a memory and a program or instruction stored on the memory and executable on the processor, the program or instruction when executed by the processor implementing the steps of the performance assessment method of a roadside algorithm as claimed in any one of claims 1 to 12.
15. A readable storage medium, characterized in that it has stored thereon a program or instructions which, when executed by a processor, implement the steps of the performance evaluation method of a roadside algorithm as claimed in any one of claims 1 to 12.

Description

Performance evaluation method and device for road side algorithm, electronic equipment and storage medium Technical Field The application belongs to the technical field of vehicle-road cooperation, and particularly relates to a performance evaluation method and device of a road side algorithm, electronic equipment and a storage medium. Background With the rapid development of automatic driving and intelligent traffic systems, the requirements for road side algorithm performance evaluation are increasingly urgent. In the prior art, the performance evaluation method of the path algorithm is mainly focused on verification of a single index. For example, positioning accuracy evaluation, comparison verification through precision positioning equipment such as RTK, target detection accuracy, recall rate and accuracy rate of a detection algorithm through manual annotation verification, track smoothness and simple evaluation based on continuity of track points. Most evaluation methods in the prior art only pay attention to a single technical index, so that evaluation dimensions are single, and comprehensive performances of road side algorithms in complex traffic environments cannot be comprehensively reflected. Disclosure of Invention The embodiment of the application aims to provide a performance evaluation method, a performance evaluation device, electronic equipment and a storage medium of a road side algorithm, which can solve the problem that the evaluation dimension is single and the comprehensive performance of the road side algorithm in a complex traffic environment cannot be comprehensively reflected. In order to solve the technical problems, the application is realized as follows: In a first aspect, an embodiment of the present application provides a performance evaluation method for a road side algorithm, where the method includes: Acquiring a track sequence output by a road side algorithm, wherein each track point in the track sequence comprises position information and kinematic characteristics; Mutation detection is carried out on the kinematic characteristics of adjacent track points in the track sequence, so that a mutation detection result is obtained; according to the position information and the kinematic characteristics, carrying out anomaly detection on each track point in the track sequence to obtain an anomaly detection result; And determining a performance evaluation result of the road side algorithm according to the mutation detection result and the abnormality detection result. In a second aspect, an embodiment of the present application provides a performance evaluation apparatus for a roadside algorithm, including: The track acquisition module is used for acquiring a track sequence output by the road side algorithm, and each track point in the track sequence comprises position information and kinematic characteristics; the mutation detection module is used for carrying out mutation detection on the kinematic characteristics of adjacent track points in the track sequence to obtain a mutation detection result; the anomaly detection module is used for carrying out anomaly detection on each track point in the track sequence according to the position information and the kinematic characteristics to obtain an anomaly detection result; And the evaluation result determining module is used for determining the performance evaluation result of the road side algorithm according to the mutation detection result and the abnormality detection result. In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, and a program or instruction stored on the memory and executable on the processor, the program or instruction implementing the steps of the method according to the first aspect when executed by the processor. In a fourth aspect, embodiments of the present application provide a readable storage medium having stored thereon a program or instructions which when executed by a processor perform the steps of the method according to the first aspect. In a fifth aspect, an embodiment of the present application provides a chip, where the chip includes a processor and a communication interface, where the communication interface is coupled to the processor, and where the processor is configured to execute a program or instructions to implement a method according to the first aspect. In the embodiment of the application, the track sequence output by the road side algorithm is obtained, each track point in the track sequence comprises position information and kinematic characteristics, mutation detection is carried out on the kinematic characteristics of adjacent track points in the track sequence to obtain mutation detection results, abnormality detection is carried out on each track point in the track sequence according to the position information and the kinematic characteristics to obtain abnormality detection results, and the performance evalu