Study On Terminal Sliding Mode Control Of Manipulators Based On Neural Network And Time Delay Estimation

With the development of the times and the progress of science and technology,manipulator has been widely used in the global industry and has become one of the in-dispensable technologies for the development of modern industry.The manipulator is a nonlinear system with multi-input and multi-output,strong coupling and time-varying,and its operating environment is extremely complex.Therefore,it is particularly important to design a high performance trajectory tracking control strategy for the manipulator.In this paper,the trajectory tracking control problem of manipulator with unknown information and many uncertainties as well as external disturbances is stud-ied.Based on terminal sliding mode control method,time delay estimation strategy and neural network control theory,the following studies are carried out:As a trajectory tracking control method,terminal sliding mode control is widely used in the field of manipulator control due to its finite time convergence and invari-ance under parameter changes.The delay estimation strategy has attracted the attention of many scholars because it does not require accurate information of the model and is less dependent on the model for easy application.Therefore,a terminal sliding mode control strategy based on delay estimation is proposed by combining the delay estima-tion strategy and terminal sliding mode control method.In this strategy,the time delay estimation method is used to estimate and compensate the uncertainty in the system to offset its influence.At the same time,the terminal sliding mode control is applied to track the trajectory.By introducing nonlinear terms,the fast terminal sliding mode surface function is designed to reduce the time required for the system state to converge to the equilibrium point.In addition,the robust control term is designed to overcome the chattering phenomenon of the terminal sliding mode controller.The effectiveness of the proposed method is verified by simulation.Considering the nonlinear term introduced in the controller previously designed,the function has a negative fractional power.When the parameters are not properly selected and the system error converges to zero,the singularity phenomenon will occur,that is,a signal state is infinite.In addition,the time delay estimation will lead to large error when there are short time changes in the system,which will reduce the performance of the system.To solve the above problems,a non-singular terminal slid-ing mode control strategy based on neural network is proposed.Radial neural network algorithm has some characteristics similar to delay estimation,and it can avoid the local minimum problem of the system.Considering that the traditional radial neural network adopts weight adaptation,which requires a large number of adaptations and a complex control structure,resulting in a slow solving speed,a small number of virtual auxiliary parameters are designed to replace the weight of neural network,which improves the approximation solving speed of neural network and optimizes the controller structure.Then an adaptive non-singular terminal sliding mode controller is designed based on the equivalent control principle,and a robust term is added to eliminate chattering caused by the terminal sliding mode.The closed-loop stability of the proposed controller is proved by using Lyapunov stability theory.Finally,the trajectory tracking performance of the proposed method is verified by simulation.