Improving predictive control of a mobile robot: Application of image processing and Kalman filtering

This paper discusses a control algorithm for the interception of a mobile target. The application domain is robot soccer in which the target is a ball while the interceptor is a wheeled robot. A case study of a wheeled robot approaching a target in a given direction is presented in detail. A shoot function has been developed, using the proposed algorithm, which calculates the robot wheel velocities to position the robot behind the ball. In order to improve the efficiency of the control, a scaling factor has been introduced which can be optimized so that the robot travels a shorter distance to intercept the target. Since the target is constantly moving, it is imperative that the future position of the target be predicted and used in the control algorithm to calculate the wheel velocities. Kalman filtering has proven to be an effective tool for predicting the target position at the point of interception, several frames ahead and thus improve the accuracy of interception.