Vehicle Tracking and Controlling Based on Roadside Cameras

In this paper, we design a vehicle target tracking system based on roadside cameras. This system contains vehicles and devices on roadside including computation units. The motion of vehicles is captured by cameras, and then is sent to computation units, in which the YOLO-V3 algorithm is used to identify the target vehicles. In particular, the Kalman filter is utilized to estimate the position and velocity of vehicles we are interested in. Moreover, we design a PID control law to drive the target vehicles to the designed sites. Experimental results show that the system we design can achieve vehicle target tracking by using roadside cameras.