FPGA Implementation Of Neural Network PID Motion Control Algorithm For Home Service Robot

With the continuous improvement of people's economic level and the aggravation of population aging,family service robots,as a tool to facilitate People's Daily life,have become more and more widely used.Home service robot technology is one of the key areas of national scientific and technological development.Because home service robots have a variety of uncertainties such as nonlinearity and time-varying,in addition,people have higher intelligent technical requirements for their performance,so families The control mechanism of service robots is becoming more and more complicated,and its mathematical model is also difficult to establish.Therefore,in order to improve the control accuracy of the home service robot and meet its higher requirements for stability and real-time performance in motion control,this paper studies the traditional PID closed-loop control algorithm used in the home service robot motion control system,and for the fixed invariance of the control parameters of this algorithm,and the control commands sent by the motion control system to the host computer in a specific environment,there are problems in real-time and stability,utilize a neural network with powerful nonlinear mapping capabilities and self-learning capabilities,and its field programmable device FPGA with parallel and efficient processing capabilities,an optimization algorithm that uses BP neural network to improve the traditional PID control algorithm is designed,and FPGA-based robot wheel structure chassis motion control system is implemented based on this algorithm.In this paper,based on the in-depth study of the omnidirectional mobile robot motion characteristics and the principle of neural network PID control algorithm,Simulink is used to model and simulate the designed closed-loop control algorithm;Then use the top-down modular design method of Verilog language,use FPGA to realize BP neural network PID closed loop control system,and based on the Modelsim software platform,the timing waveform analysis of each designed module is carried out.Finally,download the successfully compiled file to the FPGA main control board for physical test.This article uses the three-wheeled omnidirectional mobile robot chassis as an experimental platform,combined with the existing robot upper computer of the laboratory that can realize the following and autonomous navigation functions,the physical test of the chassis motion control system is tested.The results show,the motion control system implemented in this article improves the running speed of the chassis by 11.6% and the accuracy by 17.1%,therefore,the reliability and practicability of the BP neural network PID motion control algorithm in home service robot control are effectively verified.