Research On Control Of Types Of Robots Based On Backstepping

Robot is a kind of electromechanical integration equipment which can be achievedhumanoid operation and automatic control, repeat programming, finish all kinds ofassignments in 3d space. It is used in manufacturing firstly. With the development ofscience and technology, its application has expanded to many other fields now such asnational defense, medical treatment and public health, disaster relief and life service, etc.Since robot system has wide application prospects, researchers of the world have payedhigh attention to its control.Based on backstepping method, while combine with sliding mode variable structurecontrol, adaptive radial basis function(RBF) neural network, and dynamic surface control,this paper designs controllers for three different types of robot respectively. The specificwork is as follows:Frist, translating wheeled mobile robot’s mathematical model into a chain systemwhich has no drift and double inputs by coordinate and input transformation. Thenstabilizing the system with backstepping and sliding mode variable structure controlmethod. Input-state scaling technique is used for the design of virtual control law in thefirst two steps. In order to make the system’s last state converging in limited time, variablestructure control is adopted in step 3.A adaptive backstepping design scheme for rigid robot manipulators with parametricuncertainties is proposed based on the visual feedback and fully tuned RBF neuralnetworks. The feature extraction by the CCD camera mounted on the end effector makesthe desired position. Fully tuned RBF neural networks which different from the generalRBF neural networks are used to approximate parametric uncertainties and disturbances ofthe system. The center value and incidence of the fully tuned RBF neural networks areadjusted as well as the weight. Therefore, the approximate ability of the networks isimproved on line. It is proved by Lyapunov stability theory that all signals of the systemare bounded and the controller guarantee exponential convergence of the motion of robotmanipulators to the desired position. In this paper, practical stable control based on adaptive fuzzy dynamic surfacescheme is presented for single-link flexible robots. The unknown and time-varyingfunctions are compensated by fuzzy approximators. By using DSC design technique, theproposed control system can overcome the problem of“explosion of complexity”inherentin the backstepping design methods. Thus, the structure of the controller is simplified. Theproposed controller guarantees the practical stability of the closed-loop system.This paper has conducted simulation for the three different kinds of controller. Theresults of numerical simulation show the effectiveness of the proposed controller for thesethree different robots.