Practical Tracking Control For Multiple Classes Of Uncertain Robotic Systems

Tracking control of robotic systems has always been a hot topic in control theory and application.On one hand,strong nonlinearities and serious uncertainties are usually involved in the system.On the other hand,certain safety conditions,such as input/output constraints,should be considered to guarantee the safety of system.Then,control design for such systems are usually rather challenging,which brings many control problems of scientific and practical significance in this field to be solved.For the deficiency of the existing control theory results,both the compensation mechanism of system uncertainty and those for the input/output constraints as well as obstacle-avoidance are proposed in the paper for multiple classes of typical uncertain robotic systems,and meanwhile,new frameworks for tracking control design and performance analysis are developed which solve the control problem of secure practical tracking control of uncertain robotic systems.The main research contents are given as follows:Firstly,practical tracking control for a class of uncertain surface vessel systems is investigated.Different from the related literature where the severe assumptions on the system uncertainties and reference trajectories are severely restricted,the paper allows all the system parameters to be unknown and the disturbance are not to be smooth or have known bound,while the reference signals are only first-order continuously differentiable without their time derivatives to be available for feedback.For this,by incorporating an adaptive dynamic compensation mechanism into backstepping scheme,a new practical tracking control framework is proposed with an adaptive state feedback controller.The proposed controller guarantees that all signals of the resulting closed-loop system are bounded while the system output practically tracking the reference trajectory.Secondly,practical tracking control for a class of robotic systems with obstacleavoidance is investigated.Compared with the existing literature,besides the serious system uncertainties,the dynamic obstacles have less available information in the paper,such as their speeds are not necessarily for feedback.For this,an adaptive dynamic compensation mechanism is proposed to overcome serious system uncertainties while a barrier function is skillfully incorporated into backstepping scheme to guarantee the obstacle-avoidance performance.Then,a tracking controller is explicitly designed to ensure the obstacle-avoidance and the practical tracking.Thirdly,fault-tolerant practical tracking control is investigated for uncertain robotic system driven by DC motors with output constraint.Compared with the existing literature,besides the serious system uncertainties,the output constraint functions have less available information in the paper,such as their time derivatives are not necessarily for feedback.For this,an adaptive dynamic compensation mechanism is proposed to overcome serious system uncertainties and actuator faults while an appropriate Barrier Lyapunov function is chosen to guarantee the output constraints.Then,a tracking controller is explicitly designed to guarantee the output constraints and the practical tracking.Finally,practical tracking control via switching is investigated for a class of general uncertain nonlinear systems which represent multiple classes of robotic systems.Remarkably,the system possesses essential nonlinearities and serious uncertainties while dead-zone input and more general conditions of output constraints are considered in the paper,which make the methods in the related literature invalid.For this,a logic switching mechanism is designed to tune a pivotal controller parameter online,and hence to compensate system uncertainties,dead-zone input and output constraints.Then,a switching controller is explicitly designed which guarantees the practical tracking performance of the closed-loop system.