Multiple Object Tracking using Radar and Vision Sensor Fusion for Autonomous Vehicle

Autonomous Vehicle (AV) employs multiple sensors to sense the surroundings and take decisions accordingly. The aim of the research presented in this paper is to design a sensor fusion and tracking algorithm that tracks the objects based on data fusion of RADAR and Camera. The developed algorithm helps in not only accurately identifying the objects but also determining object's position. The algorithm is tested on a testbed vehicle by fusing RADAR and Vision sensor data. The clutters in RADAR data are detected and removed thus increasing reliability. To avoid the problem due to false data from the sensors, the sensor data is fed to an Interactive Multiple Model (IMM)filter comprising of three Kalman filter. Fusion analysis data shows that the designed IMM filter gives better results as compared to a single Kalman filter.