Motion Prediction and Risk Assessment for The Decision Making of Autonomous Vehicles

Abstract

The last few years, the automotive industry sees the Autonomous Vehicles (AV) as a great opportunity to increase comfort and road safety. One of the most challenging tasks is to detect dangerous situations and react to avoid or, at least, mitigate accidents. This requires a prediction of the evolution of the traffic surrounding the vehicle. This paper is a survey of the methods used in Automotive engineering for predicting future trajectories and collision risk assessment. models of vehicles are classified from the simplest to the more complexes. These technologies aim to improve road safety by estimating the level of dangerousness of a situation to make decision to avoid collision or mitigate its consequences.