Study Of Intelligent Control Algorithm Based On Binocular Visual Servo System

Visual system is the important mechanism for wheelchair robot to get outside information, how to improve the sensing range and performance of system is great significance for wheelchair robot's fast and accurate decision-making. Binocular visual servo system can not only expand the perception scope of visual system, getting lots of three-dimensional image information, but also has a small amount of processing for image information, so it becomes the central issue of research gradually.In recent years, intelligent methods and theory develops fast, it provides effective approaches for the solution of problems such as:object model can not be accurately obtained; Nonlinear and time-varying systems. In order to improve the fast, stable and synchronization performance of binocular visual servo system for wheelchair robot, the paper adopt related intelligent control algorithm, including fuzzy control, adaptive genetic control and BP neural network control, and the simulation of the controller designed is carried out through MATLAB.First, basic control structure of binocular visual servo system is analyzed and designed, meanwhile the rotational velocity distribution for camera is also designed, the simulation model of the system is established in MATLAB/SIMULINK, and simulation results are obtained.Second, make a study on fuzzy and genetic control algorithm, combine fuzzy control with traditional PID, and the fuzzy controller adjusts the PID parameters online and real-timely. In view of the drawbacks for fuzzy control, adaptive genetic algorithm is adopted to make an optimization for fuzzy rules and membership function of fuzzy controller. Simulation results show that compared with traditional PID, the method proposed in this paper has a strong robustness to disturbance, and speed can increase quickly, meanwhile, the synchronization performance of the two motors is also improved greatly.Finally, in order to further improve the fast tracking performance and binocular synchronization performance of binocular visual servo system, feed-forward control method is adopted, and BP neural network is used to approach the inverse model of controlled object, solving the problem that the model of feed-forward controller is hard to construct in practice. Simulation results show that the method is effective and feasible.