Collision Risk Analysis to Ensure the Safety of Autonomous Vehicle Motion Control

Abstract

This paper discusses the problem of ensuring the safe (accident-free) autonomous vehicle (AV) movement along a trajectory. The task of ensuring the safety of autonomous vehicle movement along the trajectory is formalized. The article describes a model that quantitatively assesses the level of autonomous vehicle movement safety based on the analysis of the environment map using a risk-based approach. A method for finding the optimal moment of time for an emergency stop of an AV is presented on the basis of minimizing possible damage (losses) by analyzing the change in collision risk as a safety indicator. The author has simulated a collision between an AV and an obstacle in Matlab Simulink. The probability densities of the collision risk values in the event of an AV hitting an obstacle and in the absence of an accident were constructed based on the simulation results. The author has shown the possibility of using the results of accident modeling to solve the problem of ensuring safe autonomous vehicle movement.