An Extended Kalman Filter Based Object Tracking Algorithm for Autonomous Guided Truck Using 1-Layer Laser Scanner

This paper describes an extended Kalman filter based object tracking algorithm for autonomous guided truck using 1-layer laser scanner. The 1-layer laser scanner has been used to obtain 2D cloud point data to detect the preceding object for tracking control. The object tracking algorithm proposed in this study consists of perception, decision, and control stages. In the perception stage, object’s information such as relative coordinate and yaw angle has been derived based on coordinate transformation, clustering, and state estimation algorithm using the obtained point data from laser scanner. In order to estimate object’s states such as coordinate and velocity, the extended Kalman filter has been used in this study. Based on the estimated states of the object, the desired path has been derived for calculation of steering angle. The simplified mathematical model of the truck has been derived to design optimal controller. The optimal controller designed in this study is based on the linear quadratic regulator for computing the optimal angle of steering module used for tracking. In order for reasonable performance evaluation, actual data from laser scanner and the derived mathematical model of truck have been used. The developed tracking algorithm and performance evaluation have been designed and conducted on Matlab/Simulink environment. Results of the performance evaluation show that the developed object tracking algorithm has been able to track the preceding object using 1-layer laser scanner.