Safety virtual verification method for expected function of automatic driving system

Abstract

The invention discloses a safety virtual verification method for an expected function of an automatic driving system. The method comprises the following steps: establishing a dynamic virtual world and a tested automatic driving vehicle model in a computer system; setting physical truth value data of the detection target in the virtual world; dividing a vehicle driving area in the virtual world into gridding sub-areas; starting a test, and calculating an expected KPI (Key Performance Indicator) of the detected target in each sub-region in the virtual world; calculating the KPI shown by the automatic driving system in the running process of the tested vehicle; the coordinate origin of the tested vehicle is matched into the grid covered by the tested vehicle; comparing the expected KPI of the sub-region with the KPI represented by the tested vehicle to obtain comparison data, and generating a safety analysis report; performing quantitative analysis and calculation on comparison results, and extracting a scene time slice set with design defects as an unsafe scene library for further verifying improvement measures; and repeating the steps to retest the improved automatic driving system until the safety requirement is met.

Description

technical field [0001] The invention relates to the technical field of safety verification methods for automatic driving systems, in particular to a virtual verification method for expected functional safety of automatic driving systems. Background technique [0002] The verification of the expected functional safety of autonomous driving requires a large number of tests to ensure that the system can cope with the scenarios required by the safety design, but large-scale road tests are costly and inefficient. By virtualizing the scenes of the physical world in the computer, it is an effective method to allow self-driving vehicles to conduct large-scale road tests in the virtual world to find design deficiencies. This method can improve efficiency, reduce costs, The scope of verification is expanded, but the quantitative evaluation of safety is still insufficient. Contents of the invention [0003] The technical problem to be solved by the present invention is to overcome t...

AI Technical Summary

Problems solved by technology

[0002] The verification of the expected functional safety of autonomous driving requires a large number of tests to ensure that the system can cope with the scenarios required by the safety design, but large-scale road tests are costly and inefficient

By virtualizing the scenes of the physical world in the computer, it is an effective method to allow self-driving vehicles to conduct large-scale road tests in the virtual world to find design deficiencies. This method can improve efficiency, reduce costs, Expanded scope of verification, but still insufficient in terms of quantitative assessment of safety

Images