Composite Nonlinear Feedback Control For Uncertain Robot Manipulators

Based on the robot joint trajectory tracking task,this paper studies the motion control problem of uncertain robots.Firstly,the composite nonlinear feedback(CNF)control technique of high-order multivariable saturated linear systems is studied and applied to the robot control system.Then combine the robust control with adaptive control and CNF control to design a new controller,which aims to effectively eliminate the adverse effects of uncertain factors on the robot system.Finally,the problem of the stable attraction domain of the robot system is discussed,and the uncertain CNF controller with adjustable parameters is designed.The main research contents are as follows:(1)The control characteristics of CNF control are studied,which is composed of linear feedback and nonlinear feedback: the role of linear feedback is to keep the system faster;the purpose of nonlinear feedback is to adjust the damping rate of the system to avoid overshoot.Introducing CNF control into the general robot system can prove the stability of the robot system under the CNF controller and achieve good tracking effect on the desired trajectory.(2)When the robot system is uncertain,the original CNF controller no longer has the ability to accurately track the desired trajectory.The basic feature of robust control is to use a controller with fixed structure and parameters to ensure that the design requirements can be achieved even if the uncertainty affects the performance quality of the system.The combination of robust control and CNF control is designed to preserve the impact of CNF control on the system to compensate for the effects of eliminating uncertainty on the system and to achieve effective control of the uncertain CNV controller for uncertain robot systems.(3)Using the adaptive fuzzy control system with fuzzy generator and fuzzy eliminator to estimate the uncertainty online,and use the estimation result as the compensation term of CNF controller,design a CNF controller based on adaptive fuzzy compensation.It is used to reduce the impact of uncertainty on the system and to make the system more stable and robust.(4)Using CNF control can enhance the control capability of the driver and improve the dynamic characteristics of the system.Apply CNF to the motion control of uncertain robots,and combine with the adaptive RBF neural network to study the adaptive RBF neural network based on CNF.The control performance of the networkcontroller.The adaptive RBF neural network controller is independent of the model-based CNF controller,and its cut-in does not affect the original control system design.(5)In order to eliminate the influence of uncertainty disturbance on the robot system,it is proposed to reconstruct the disturbance link using the observer to compensate the disturbance and improve the tracking performance of the robot system.That is,the interference compensation and interference estimation items are added to the original CNF control structure,and the robot CNF controller based on the disturbance observer is designed.(6)The stable region of the robot CNF controller is local in nature,and the attraction domain estimation of the robot system is also conservative.Therefore,we can learn the gain parameter adjustment method of high and low gain theory,and introduce the adjustable CNF controller with adjustable parameters by introducing adjustable parameters to the linear gain and nonlinear gain.It can not only expand the attraction domain of the closed-loop system,but also have respond to smooth dynamics and other advantages.