A method, device, electronic equipment and storage medium for determining a virtual test scene

Abstract

The present application relates to the field of autonomous driving technology, and in particular to a method, device, electronic device and storage medium for determining a virtual test scene, so as to improve the construction efficiency and authenticity of the virtual test scene of the autonomous driving vehicle, wherein the method includes: obtaining the input Scene configuration parameters, and determine the test map according to the map information in the scene configuration parameters; configure the test map to generate a virtual test scene corresponding to the test map according to the automatic driving vehicle configuration information in the scene configuration parameters, wherein the virtual test scene includes at least A self-driving vehicle and at least one scene element; performing a virtual test on the self-driving vehicle according to the virtual test scene; and determining a virtual test scene where abnormalities occur during the virtual test as a virtual test scene of the self-driving vehicle. Since the application can automatically generate large batches of simulation test scenarios, the construction efficiency and authenticity of autonomous driving scenarios are improved.

Description

technical field [0001] The present application relates to computer technology, in particular to the field of automatic driving technology, and provides a method, device, electronic equipment and storage medium for determining a virtual test scene. Background technique [0002] Self-driving vehicles perceive the surrounding environment of the vehicle through the on-board sensing system, and control the steering and speed of the vehicle according to the road information obtained by the perception, other vehicle position information, and obstacle information, so that the vehicle can safely and reliably travel on the road. drive on. [0003] In related technologies, the self-driving test scene is constructed by simulation software and then the self-driving vehicle is simulated and tested. A common way to construct the self-driving test scene is: a custom scene based on a scene editor, traditional simulation software such as VTD (Virtual Test Drive, virtual test drive), PreScan,...

AI Technical Summary

Problems solved by technology

[0004] However, the scene editor relies entirely on user input to generate the scene, such as figure 1 As shown, it is more convenient for users to select pedestrians, obstacles, vehicles, etc. to generate a small number of scenes, but autonomous driving above L3 (Level 3, Level 3) requires a scene library in units of tens of thousands, and the huge workload leads to Manually editing scenes with the scene editor became an unrealistic option

At the same time, the behavior of artificially defined scene elements is not realistic enough, which has a negative effect on the training results of the method

Images