Trajectory Tracking And Formation Control Of Wheeled Robot Based On Model Predictive Control

With the development of robot kinematics,control theory and computer hardware technology,the research on robot control has become a hot research topic with wide attention at home and abroad.At present,robots have many applications in various fields,which can effectively liberate productivity and show great superiority in assisting human work.The stabilization,trajectory tracking and formation control of mobile robots are the basic research contents of robot control,and the discussion of these aspects is of great guiding significance to the further development and improvement of robotics.For the control of wheeled mobile robots,the traditional control methods don't take the constraints of wheeled mobile robots into account.Therefore,the control method based on model prediction controller is adopted and the constraint problem is solved for differential wheeled mobile robot and omnidirectional wheeled mobile robot respectively.Then the quadratic programming problem transformed from model predictive control is solved online,and the optimal control input is obtained by iterative solution.The stabilization control of a single differential wheeled mobile robot is studied firstly.The stabilization control chain system is established and the model prediction controller is used to realize the stabilization of the robot.The trajectory tracking of the differential wheeled mobile robot is further studied and the kinematic error model of the trajectory tracking is established.Aiming at the problem of kinematics constraint which is only considered in the general control method,the model predictive control and the adaptive control method are combined to consider the dynamic constraint of the robot.Secondly,due to the unknown interference in the actual control system,the state feedback controller is designed based on the model prediction controller,so it compensates for the unknown disturbance.Then,the model predictive control is applied to omnidirectional wheeled mobile robot,and the trajectory tracking error model is proposed.Considering the speed constraint of the robot,the trajectory tracking control is realized.Finally,based on the research of a single wheeled mobile robot,a leader-follower robot formation model is established to realize formation control of omnidirectional wheeled mobile robots.And the formation system is controlled by the model predictive control method to achieve the desired relative position relation of robot.A delay neural network is proposed to solve the quadratic programming problem derived from the model predictive control approach.The neural network can quickly solve the constrained quadratic programming problem in real time and obtain the optimal solution.Meanwhile,primal-dual neural network is also used to solve quadratic programming problems.The two are compared with the Lagrange neural network.The model predictive control is adopted to realize the motion control of wheeled mobile robot,including the predictive stabilization control based on delay neural network,the trajectory tracking control based on kinematics and dynamics,and the formation control.Finally,the effectiveness of the model predictive control for robot stabilization,trajectory tracking and formation control is verified by simulation.