An Active Compensation System for Robot Vision in Rough Environment Based on Bionic

Abstract—Motion targets tracking techniques is one of the most important aspects in the fields of mobile robot research under unknown environments. To solve the problem of unstable vision resulted from the uncertainties, a method of active compensation for robot visual was proposed based on the principle of bionic eye movements. According to oculomotor neural circuits, an adaptive oculomotor control model of eye movements was established, and the model includes that VOR, smooth purisuit and compound eye movements. In order to verify the model’s performance, some simulation experiments ware conducted in different environment. Simulation results show that the model can be active compensation of visual error caused by the dynamic variation of the robot attitude and the tracking target location. Finally, physical robot experiments results also confirm the effectiveness of the control model. Compared with the conventional camera, this new one can solve the problem of unstable vision. Index Terms—bionic eye movements, oculomotor control modeling, oculomotor neural circuits, visual compensation, bionic mechanical pan-tilt camera. I