A novel method for vision tracking with line of sight control, using a fuzzy logic controller and Euler angle orientation in the feedforward loop

Abstract

This paper presents a novel method for vision tracking with line of sight control. The line of sight control for the vision tracking is an important function for various robot applications. However, when a robot is moving fast, it is difficult to keep the camera's sight towards a specific target, using only a vision feedback controller, due to the limitation in vision processing capabilities. In the proposed method, a feedforward controller based on robot position information is implemented to improve the tracking performance. The translational motion of the robot is estimated using a fuzzy logic controller and a slip detector. In addition, an Euler angle expression is used to calculate the robot orientation in 3-D space. Then, a vision-based feedback controller is added to compensate for the feedforward targeting error. To evaluate the performances, the developed vision tracking system is implemented on a mobile robot, and various experiments are performed. Experimental results show excellent tracking and recognition performances.