Automatic driving vehicle test simulation scene generation method

Abstract

The invention discloses an automatic driving vehicle test simulation scene generation method, and belongs to the field of vehicle testing. Real road network information in a certain area is extracted based on a high-precision map, supplementary definition parameters of related roads and environments are generated based on a Monte Carlo method, and the supplementary definition parameters and the real road network information jointly form an automatic driving vehicle test static simulation scene. An initial state of a plurality of manual driving vehicles is generated based on a Monte Carlo method, and each manual driving vehicle is endowed with a driver module, so that the manual driving vehicle has basic logic judgment capability and driving behavior capability in vehicle driving, and an automatic driving vehicle test dynamic simulation scene is perfected. According to the automatic driving vehicle test simulation scene generation method developed by the invention, factors such as road conditions, traffic conditions, road identifiers and weather are comprehensively considered, rapid establishment of simulation and test typical scenes is realized, the problems of low test feasibility, long time consumption and high cost during automatic driving vehicle test are solved, and resource waste is avoided.

Description

technical field [0001] The invention belongs to the field of automatic driving test, and in particular relates to a method for generating a test simulation scene of an automatic driving vehicle. Background technique [0002] SAE L2 / L3 autonomous driving technology is gradually reflected in mass-produced models, and some OEMs and autonomous driving core technology suppliers are conducting research on forward-looking technology and mass production of L4 / L5 autonomous driving. However, there is a huge difference between allowing a small number of self-driving vehicles to drive in a relatively closed environment and allowing them to drive autonomously in a real environment. Before mass production, self-driving vehicles need to conduct a large number of tests to ensure their safety. safety and reliability. At present, car companies at home and abroad mostly conduct real vehicle tests in closed / semi-closed loop sites, which will undoubtedly consume a lot of time, manpower, and ma...

AI Technical Summary

Problems solved by technology

However, there is a huge difference between allowing a small number of self-driving vehicles to drive in a relatively closed environment and allowing them to drive autonomously in a real environment. Before mass production, self-driving vehicles need to conduct a large number of tests to ensure their safety. safety and reliability

At present, domestic and foreign car companies mostly conduct real vehicle tests in closed / semi-closed loop sites, which will undoubtedly consume a lot of time, manpower, and material resources, and some typical scenarios cannot be reproduced

Some car companies also use SIL (Software in Loop), HIL (Hardware in Loop), and VIL (Vehicle in Loop) methods to artificially build typical scenarios in the scenario simulation software, but this does not satisfy the requirements of unlimited and scalable autonomous driving tests. Requirements for precision, mass production and automation

Images