Software-in-the-loop testing method for simulation-based automatic driving tracking algorithm

Abstract

The invention discloses a software-in-the-loop testing method of a simulation-based automatic driving tracking algorithm, comprising: establishing a simulation scene of the tracking algorithm, generating simulation data of the simulation scene; visualizing the simulation data, and displaying the test path and the actual track The difference in; the tracking function score is given according to the degree of track coincidence. Provide a real-time simulation test environment for the automatic driving tracking algorithm, obtain real-time vehicle speed and trajectory during the simulation process, improve the testing efficiency of the automatic tracking algorithm, reduce test energy consumption, and obtain comprehensive evaluation results.

Description

technical field [0001] The invention relates to the technical field of automatic driving testing, in particular to a software-in-the-loop testing method based on a simulation-based automatic driving tracking algorithm. Background technique [0002] In recent years, unmanned driving-related technologies have received widespread attention. Since unmanned vehicles often face high risks when testing, hardware reliability, algorithm robustness, etc., the generalization ability for complex environments will have a great impact on the final effect of autonomous driving algorithms. . At the same time, since the autonomous driving algorithm contains multiple modules, the compatibility between each module will also affect the actual use of the algorithm. In addition, testing unmanned driving algorithms in some dangerous scenarios is very dangerous and will greatly increase the cost of testing. Therefore, before the actual road test, it is necessary to establish a simulator test pla...

AI Technical Summary

Problems solved by technology

[0003] At present, the existing unmanned vehicle simulation test methods are limited to using simulation data calculation results to evaluate and analyze the performance of unmanned driving test algorithms. Because the simulation data is not updated and interacted in real time, the test efficiency is low and the evaluation results are not enough. Accurate and comprehensive, the simulation effect is relatively poor

It uses the simulated vehicle trajectory in the vehicle motion simulator to evaluate the performance of the unmanned driving algorithm. This single trajectory evaluation cannot reflect the real speed change of the vehicle during the driving process.

Images