Visual Tracking of Unknown Moving Object by Adaptive Binocular Visual Servoing

Visual tracking of moving objects is one of the most fundamental capabilities for intelligent robots to accomplish the given task in unknown, dynamic environments. Visual servoing which utilizes direct feedback from image features to motion control has been used to realize such capability. However, the conventional visual servoing needs much knowledge on the system parameters such as kinematics and optic ones and/or on target objects. Adaptive visual servoing [1] was proposed to remedy such a limitation by adopting online estimation of image Jacobian. However, it is limited to track stationary objects or moving ones of which motion is known in advance. This paper proposes an extension of the adaptive visual servoing for unknown moving object tracking. The method utilizes binocular stereo vision system, but does not need the knowledge of camera parameters. Only one assumption is that the system needs stationary references in the both images by which the system can predict the motion of unknown moving objects. The experimental results are shown and a discussion is given.