Study On Movement Control For Bionic Eye Based On Human Eyes Movement Mechanism

It is important for eyes to sense the outside world and interact with the environment. Eyes have extremely complex and subtle nerve control mechanisms in sensing, recognising and tracking the target. They can help people to cognitive the environment and objects more quickly and accurately. The rapid development of machine vision makes robots have the ability to interact with the environment. But the gap to get to advanced human visual system is big. Therefore, it is of great research significance to get a deep understanding of the control mechanism of the human eye system and merge it with machine vision, control science and other science technologies for increasing the peroception ability of robots. This thesis mainly studies the bionic eye movement control algorithm. The main works can be concluded as follows:Firstly, we analyze the research background and significance for bionic eyes, review and summarize the current research, point out the main problems, and give the main works and framework of this work.Secondly, the structure and characteristics of eye balls are studied. The principle of image acquisiton is studied. And the oculomotor nerve control mechanism is studied.The relationship between the eye movement and biommimetic visual perception is disscussed.Thirdly, the control mechanism of the smooth pursuit eye movement is studied. A smooth pursuit eye movement model is built. The model is proposed based on Kalman filter and Iterative Learning Control(ILC) algorithm. Due to the fact that eyes are easily disturbed by the noise. Kalman filter is used to filter noise. Considering ILC can track periodic targets effectively, we apply ILC to the control of eye movement. By learning the previous information of the target motion, the optimal control command for the current cycle is obtained. The time delay problem caused by image processing is solved. Moreover, a tuning method is used for optimal control gains of ILC, increasing the adaptability of the proposed model. The superior performance of the proposed control method is verified by simulation experiments.Fourthly, a bionic eye hardware platform is bulit. A driving circuit is designed to improve the hardware platform. The hardware platform is integrated by a box. PTZ camera is used as the bionic eye hardware platform and some experiments have been done on it. In order to control the movement of a target, a sliding rail is built. The control performance of the ILC algorithm is verified by the sliding rail and PTZ camera.Finally, we summarize the work of this paper and give the future research direction.