Pan/Tilt Camera Control for Vision Tracking System Based on the Robot Motion and Vision Information

Abstract

AbstractThis paper presents a vision tracking system, based on the robot motion and vision information. For mobile robots, it is difficult to collect continuous vision information while they are in motion due to the unstable vision information. To solve this problem, the proposed vision tracking system estimates the robot position relative to a target and rotates a camera towards the target based on the estimated position information. This concept is derived from the human eye reflex mechanisms, known as the Vestibulo-Ocular Reflex (VOR) and the Opto-Kinetic Reflex (OKR). In the proposed vision tracking system, the VOR concept for compensating the head motion is realized by the feedforward control using the robot motion information from a 3-axis gyroscope and wheel encoders. To realize the OKR concept, targeting errors are periodically compensated by using the vision feedback information. An actuation module consists of pan/tilt motion motors and the camera which performs the pan and tilt functions to locate a target at the center of an image plane. The proposed vision tracking system is integrated with a two-wheeled robot. The performance of the proposed system is evaluated by extensive experiments. The proposed system shows a significant improvement in tracking performance with small targeting error. Also, the necessities of the position feedforward and vision feedback controls are verified by the experiments for the illumination change condition and the static tilting motion.