Command Filter Backstep Impedance Control For Multi-joint Manipulator System

Impedance control has the advantages of less computational effort,strong anti-disturbance capability and easy force control,and has become a commonly used method for force/position control of multi-joint manipulator.However,the multi-joint manipulator is a highly nonlinear system with strong coupling,difficult to obtain accurate model parameters and unknown external disturbances.In the face of these problems,designing an effective controller to make the multi-joint manipulator system in an unknown external environment achieve a better position tracking effect and ensure a good impedance control performance,and avoid damage caused by large collision forces between the multi-joint manipulator and the contact object,which is research significance work.In order to deal with problems of difficulty to obtain accurate model parameters and unknown external disturbances in the multi-joint manipulator system by combining command filtered technique,adaptive neural network control technique,disturbance observer technique and finite-time control technique with backstepping method,and designs an adaptive neural network command filtered backstepping impedance control method,a disturbance observer-based adaptive neural network command filtered impedance control method,and a disturbance observer-based neural network finite-time command filtered impedance control method are designed to guarantee better position tracking effect and impedance control performance of the multi-joint manipulator system when contacting with the unknown external environment.The main research results of this thesis are summarized as follows:Firstly,the adaptive neural network command filtered backstepping impedance control strategy for the multi-joint manipulator system is studied to address the problem of difficult to obtain the model parameters accurately in the multi-joint manipulator system.Design of impedance controller by backstepping method based on dynamics model of multi-joint manipulator system in Cartesian coordinates.The "computational complexity" and "singularity" problems in the traditional backstepping design process are solved by combining the command filtered technique,and the filter error compensation mechanism is introduced into the controller design to eliminate the effect of filter error.Then,the adaptive neural network technique is used to deal with the uncertain nonlinear function terms in the multi-joint manipulator system.Finally,the stability of the control system is determined based on Lyapunov's theorem,and the effectiveness of the designed adaptive neural network command filter backstepping impedance control strategy for the multi-joint manipulator system is verified through simulation experiments.Secondly,a disturbance observer-based adaptive neural network command filtered impedance control strategy is investigated to address the problem of unknown external disturbances in the multi-joint manipulator system.By combining the disturbance observer and the backstepping method,the disturbance observer is used to estimate the unknown external disturbances and avoid the need for additional force sensors to estimate the external unknown disturbances in the multi-joint manipulator system,which saves the cost of the multi-joint manipulator application.Finally,the stability of the control system is determined based on Lyapunov's theorem,and the effectiveness of the method is verified by simulation experiments.The simulation results demonstrate that the disturbance observer-based adaptive neural network command filtered impedance control algorithm can well estimate the unknown external disturbance in the multi-joint manipulator system,thus improving the tracking accuracy of the control system and guaranteeing the safety of the physical interaction of the multi-joint manipulator with the unknown external environment in the case of unknown external disturbance.Thirdly,in order to ensure the safety of the multi-joint manipulator in contact with the unknown external environment,it is necessary to ensure that the position tracking error of the multi-joint manipulator converges in finite time,the disturbance observer-based neural network finite-time command filtered impedance control strategy is investigated.The finite-time command filtered technique is combined with the backstepping method,and a new filter error compensation mechanism is designed to ensure that the filter error compensation can converge in finite time.Then,the adaptive neural network impedance control is combined with a nonlinear disturbance observer to solve the problem of uncertain nonlinear function terms and unknown external disturbances in the multi-joint manipulator system,and the finite-time control is introduced to ensure the multi-joint manipulator position tracking error can converge in finite time.Finally,the stability of the control system is determined according to Lyapunov's theorem,and the effectiveness and superiority of the method are verified by simulation experiments.From the conclusions of the simulation experiments,it can be seen that the designed control strategy has faster convergence speed and smaller control accuracy than the other proposed controls.This is conducive to the better physical interaction of the multi-joint manipulator with the unknown external environment.