Adaptive Neural Network Control Of Uncertain Robot Systems With State Feedback And Output Feedback Methods

In recent years,the control design and stability analysis for a uncertain robot have received considerable attention.In practical control systems,unknown nonlinearities,such as input deadzone,state constraint,are the most common actuator characters we have to deal with.It is usually known as a static nonlinearity which is insensitive to small signals.In many cases,neglecting the nonlinearity effects may lead to undesired performances such as excessive steady state error,poor transient response and large overshoot.Therefore,it is necessary for us to consider the possible nonlinearity effect for a uncertain robot.In this paper,we aim to solve the control problem for a uncertain robot system,under the condition of different non-linearity.We design a Lyapunov function to ensure the stability of the whole system.Besides,the uncertainty can be handled by using the learning ability from neural network structures.Both full state feedback control and output feedback control methods are designed to control the uncertain robot system.Since the robot system has a very complex characteristics,it is difficult to apply tools to the robot system design and analysis effectively.Although the robot system is complex and challenging,the control design of robots has gained significant attention,due to the highly coupled nonlinear dynamics and great capability in performing complex and complicated tasks.However,many common issue in robot control is still not addressed.Therefore,the problem for robotic control needs to be further developed.First of all,there is no uniform and applicable control strategies for the robot system.Thus,it is generally not possible to obtain a complete solution for the robotic control,which makes the robotic control focus on the qualitative properties of the time domain response,such as stability and other aspects of the study.Next,the effect of input deadzone is analyzed in the following section.Adaptive neural network-based controller is proposed to handle the effect of input deadzone,and the output of the system can achieve the desired reference.All signals in the closed-loop system is guaranteed to be bounded.Besides,simulation examples are carried out to verify the effectiveness of the proposed controller.Then,we focus on the uncertain robotic systems with input saturation.The negative effect of input saturation is analyzed and a system is proposed to handle the effect of the input saturation.Finally,aiming at the robotic system with output constraint,an adaptive neural network controller is proposed to deal with the effect of output constraint.Barrier Lyapunov functions are proposed to design an adaptive controller and all signals in the closed-loop system can be guaranteed to be bounded.For our paper,four types of systems are considered.First,the theoretical research is studied of a uncertain robot and an adaptive controller is proposed to ensure the stability of the whole system.Next,we focus on different non-linearity,such as input deadzone,input saturation and output constraint.The method of Lyapunov stability of the closed-loop system is analyzed and the system has a nice performance to track the desired trajectory.The simulation part is carried out to verify the feasibility and effectiveness of the proposed control strategies.