Visual Servo Application Using Model Predictive Control (MPC) Method on Pan-tilt Camera Platform

Abstract

Visual servoing is a technique which use feedback information extracted from a visual sensor to control the motion of a robot. Combined with an image processing algorithm, visual information from a camera becomes a visual sensor which provides feedback or measurement value i.e. center point coordinate value from detected face. The face tracking system with pan-tilt camera can be utilized for many applications. The 2 DOF (degree of freedom) camera platform being used consists of 2 servo motors which moves according to control signal calculated using a control method such that the system will be able to keep the tracked face at the center of the camera frame. In this research, Model Predictive Control (MPC) method is implemented to build a controller, in which each movement of pan and tilt servo camera is controlled by control signals calculated with MPC method. MPC application is chosen so the system can produce an optimal control signal while considering input signal saturation aspect. Optimal control signal provides by the MPC controller is produced through optimizing the system performance index. The result showed that MPC controller successfully implemented in pan-tilt camera platform. The MPC control signals deliver a good control pan and tilt system response with rise time < 1 second and overshoot < 15%.