The Research Of Multi-Robots System Structure And Trajectory Tracking Control

Along with the steadily maturity and development of computer technology, electronics engineering,control theory,artificial intelligence theory,sensor and so on, the research of robotics which formed by multi-disciplinary is also into a brand new stage.The multi-robot system which has the advantages of fault tolerant capability and high robustness can complete the mission with high efficiency and reliably in complex condition,thus it has been widely applied,and also become a hot research spot in robots.In this thesis,we researched on the architecture of multi-robot and trajectory tracking control in theory and application.Firstly,in this thesis,we analyzed and researched some intelligent robot's system structure home and abroad at present,summarized the characteristics of these robot's system structure,combined with the problem of trajectory tracking of mobile robot,and then designed a kind of layered system structure base on multi-agent system.The every intelligent agent in the multi-robot system needs to communication with each other to complete the trajectory tracking cooperatively.On the foundation of referring the past multi-robot communication technology,we use C/S module designed the multi-mobile-robot communication system,to make sure the data exchange in time and exactly.Also we make a assignment analysis aiming at trajectory tracking problem to ensure each intelligent agent works cooperatively.Aiming at the control method for trajectory tracking control,a fuzzy control system based on the principles of fuzzy control for robot trajectory tracking is structured,and the system is simulated.Not depending on mathematical model,Fuzzy logic control can effectively overcome the nonlinear and uncertainty of the controlled object.But both the difficulty in building the fuzzy rules and the inconvenience in adjusting the affirmed rules will affect the control effect.Based on the BP algorithm of neural network,combining the fuzzy control and neural network,the neural network in this thesis has the ability of fuzzy reasoning and summary,using simulation fuzzy reasoning.Thanks to the self-learning ability,the inductive reasoning method of neural network in practical control process can continuously be revised,and thanks to the clear physical meaning of the structure,we can analyze and set the structure design and weight initialization.Adopting Gaussian function as fuzzy membership functions,the thesis constructs the fuzzy Gaussian function(RBF) neural network.A trajectory tracking simulation experiment is executed based on this fuzzy neural network control system,and the results show that this method is efficient and effective.Finally,with the help of the WiRobotX80 robot platform software,this thesis design and finish the robotic tracking system in VC++ environment,and the further analyses and discussion of experiment results and data proved the practicability, accuracy and robustness of this multi-robot system structure and trajectory tracking control law.