Research On Intelligent Backstepping Sliding Mode Control Of Nonlinear Robots

Robot consists of mechanical body,controller,servo driving system andsensor device,which has characteristics of human-simulated operation, auto-matic control and reprogrammability.It is an electromechanical equipmentthat can complete various operations in three dimensions.The use of robotshas become increasingly prevalent and far reaching.Robots almost can do allof what the human do. Robots are playing a more and more important role indaily life,industrial-agricultural production,ocean probe and spaceresearch.Multilink robot is also called mechanical arm or mechanical hand.Itis connection of rigid bodys with an end fixed which characterized by spacemovement. Based on the movement,Multilink robot is a complicated MIMOnonlinear system and has the nonlinear dynamics characteristics ofnonlinearity,time varying and coupling.The aim of the control strategy is tomake the output of the rigid robot tracking a desired trajectory in perfectdynamic quality.For tracking control of multilink robot manipulators withmodeling error and external disturbance,Now the following popular methodsare used:(1)PID Control;(2)Adaptive Control;(3)Robust Control;(4)variablestructure control;(5)intelligent control;(6)Backstepping Control.This paperput forward controlling methods,which combine backstepping sliding-modecontrol,fuzzy control and neural network.Main achievements are given asfollows in this paper.(1)For tracking a desired trajectory of multilink robot manipulators withmodeling error and external disturbance,Backstepping linear sliding modecontrol,backstepping nonlinear sliding mode control and backsteppingquasi-sliding mode control are studied respectively.The system stability isproved by Lyapunov Principle.The traditional linear sliding surface is used inbackstepping linear sliding mode control which can’t converge to itsequilibrium point in finite time.A terminal sliding mode (TSM) surface is used in backstepping nonlinear sliding mode control.Compared with linearhyperplane-based sliding modes, TSM offers the superior property of finitetime convergence.Based on backstepping sliding mode control,backsteppingquasi-sliding mode Control is obtained by converting sign function intosaturation funtion.Finally,simulation results verify the validity of the controlschemes as above.(2)For tracking a desired trajectory of multilink robot manipulators withmodeling error and external disturbance,we studied new control methods inthis paper by combining fuzzy control,backstepping control and sliding modecontrol as follows:1.Global PID Fuzzy Sliding Mode Control;2.PIDBackstepping and adaptive Fuzzy Sliding Mode Control;3.BacksteppingSliding Mode Control with Fuzzy Compensation;4.Nonsingular TerminalFuzzy Sliding Mode Control Based on Backstepping.The stability of controlsystem is analyzed by Lyapunov Principle.A properly adaptive fuzzycontroller is designed to estimate uncertain upper boundary on line and trackthe modelling error and disturbance automatically in the first twomethods.Fuzzy compensation is designed to compensate the modelling errorand external disturbance in the backstepping sliding mode control with fuzzycompensation.So it weaken the disturbance and reduces chattering. Accordingto sliding mode control theory, backstepping is used to design nonsingularterminal sliding mode controller.So it overcomes disadvantage of singularpoint In order to reduces chattering a proper fuzzy controller is designed toestimate uncertain upper boundary on line and track the modelling error anddisturbance automatically. Finally,simulation results verify the validity of thecontrol schemes as above.(3)For tracking a desired trajectory of multilink robot manipulatorswith modeling error and external disturbance,new control methods werestudied in this paper by combining backstepping control, sliding mode controland neural network control as follows:1.RBF Neural Network Sliding-modeControl;2.Global PID Neural Network Sliding Mode Control;3.IntegralBackstepping Sliding Mode Control;4.Backstepping Nonsingular Terminal Neural Network Sliding Mode Control;5.Backstepping Nonsingular FastTerminal Neural Network Sliding Mode Control.The system stability isproved by Lyapunov Principle.A switching function is used as input of RBFneural network in the RBF neural network sliding mode control. RBF neuralnetwork is treated as sliding mode controller directly.So MIMO neuralnetwork sliding mode control is implemented.Global PID sliding modesurface is designed in the global PID neural network sliding modecontrol.RBF neural network is used to adjust switching gain.Controllerparameters are tuned to estimate uncertain upper boundary on line.Sochattering is reduced.In order to estimate uncertain upperboundary,Self-adaptation principle of weight is designed in the integralbackstepping sliding mode control. Nonlinear terminal sliding mode surface isused in the backstepping nonsingular terminal neural network sliding modecontrol.RBF neural network sliding mode controller is designed to estimateuncertain upper boundary to reduce chattering.In order to further minimizeconvergence time, we design fast terminal neural network sliding modecontrol.Finally,simulation results verify the validity of the control schemes asabove.