Intelligent Control Of Trajectory Tracking For Four-wheeled Omnidirectional Mobile Robot

Omnidirectional mobile robot is widely used in industrial control owing to its flexible movement,walking speed,mobility and other outstanding advantages.The four-wheeled omnidirectional mobile robot with complete constraints belongs to the nonlinear system of multi-input and multi-output,therefore,when the trajectory tracking control is carried out,the deviation between the actual trajectory and the expected trajectory is likely to be large,therefore,the effect of trajectory tracking is affected and the rapidity of trajectory tracking is reduced.Aiming at the trajectory tracking control of four-wheeled omnidirectional mobile robot,this paper mainly completed the following research work:(1)A four wheel diamond mounted robot structure is adopted,and the kinematics model of the robot is established.Then by respectively using the second Lagrange equation and the “quasi-equivalence” of the double closed loop speed control of DC motor model,the dynamic model of the robot is established,and the characteristics of the dynamic model obtained by different methods are analyzed at the same time.(2)For the trajectory tracking problem of the four-wheeled omnidirectional mobile robot,the proportion control method based on neural network,Backstepping control method,multi-mode control method is applied respectively,and the corresponding controller is designed.A kind of Human-Simulated Intelligent Controller based on neural network is put forward and designed at last.The Human-Simulated Intelligent Controller based on neural network is composed of dynamic feature perception and the control coordination layer,the parameter adjustment layer and the underlying execution layer.First,the dynamic feature perception gets the feature information and confirms the feature mode of system.Then the control coordination layer completes the qualitative modal mapping to the underlying execution layer and the quantitative error judgment to the parameter adjustment layer based on the feature mode,the parameter adjustment layer completes the parameter adjustment by using the multi-neural network according to the quantitative judgment results,and the trajectory tracking control of four-wheeled omnidirectional mobile robot is completed by the underlying execution layer according to the results of modal mapping and parameter adjustment.The controller can achieve trajectory tracking control of four-wheel omnidirectional mobile robot by combining qualitative mapping,quantitative judgment and parameter adjustment for different tracking errors.(3)The MATLAB simulation platform for trajectory tracking of the four-wheeled omnidirectional mobile robot is built.The simulation effects without disturbance and with disturbance are respectively compared.And the robot physical platform is built and the relatively experiment is conducted.Both simulation and real experiment results show that the Human-Simulated Intelligent Controller based on neural network proposed in this paper has better control effect,that is the tracking speed is faster and the corresponding tracking error is smaller,meanwhile,the system robustness is better.