Radar and Camera Fusion based Moving Obstacle Tracking for Automated Vehicles

Abstract

In this paper, a multi-sensor fusion based environment perception architecture for ground unmanned vehicles is proposed. The target-level multi-sensor fusion technology is presented to take advantages of camera and millimeter wave (MMW) radar in target perception. On this basis, a multi-target tracking model is designed to solve the problems of alignment, association, uncertainty, as well as the elimination of false data. In order to verify the stability and real-time performance of the proposed algorithm, a real vehicle test was implemented according to the statistical data and relevant indicators. The results show that the proposed algorithm can effectively perceive and track multiple obstacles in real scene.