Pan-tilt Control Method Applied to Mobile Robots

Abstract

This paper presents a pan-tilt control method based on multi-sensor fusion, PSO and BP neural network. The traditional PID control algorithm has the disadvantage of difficult parameter tuning, and it is difficult to realize the effective control of complex nonlinear system. BP neural network is used to set the PID three parameters, but the initial neural network is easy to fall into local optimal and slow convergence. In this paper, PSO algorithm is further used to set the initial weight of BP neural network, which can effectively improve the global search ability and convergence speed of BP neural network. The pan-tilt is modeled and the experiment is designed to compared with the traditional digital PID control algorithm, BP neural network PID control algorithm and PSO-BP PID control algorithm, it is proved that the control effect of PSO-BP PID control algorithm is obviously better than the former two.