Adaptive Intelligent Control Of Flexible-Joint Robots

The manipulators are mechanical devices that can imitate human arm movements and replace humans to complete some difficult,complex,and repetitive tasks.To meet the ever-increasing demands of work,manipulators with flexible joints have attracted a lot of attention.Due to the existence of this special flexibility,flexible-joint robots have more advantages in load-to-weight ratio,flexibility,safety,and energy consumption,so they are widely used in manufacturing,medical rehabilitation,aerospace,deep-sea exploration,etc.However,the flexible-joint robot not only has the same characteristics of high nonlinearity,uncertainty,and strong coupling as the traditional rigid manipulator,but also doubles the number of variables due to the existence of the flexible structure.Consequently,the comprehensive design of such systems is of practical significance,but also more complex and challenging.Under the development trend of lightweight,flexibility,networking,and intelligence,it is particularly important to design the efficient and reliable intelligent control scheme for flexible-joint robots.Therefore,this thesis mainly concentrates on the trajectory tracking control of flexible-joint robots.Meanwhile,the intelligent control design schemes that can reduce the communication burden and improve work efficiency are proposed by using the backstepping design as a framework and combining adaptive control and fuzzy logic systems/neural networks.The main work is as follows:(1)A finite-time adaptive tracking control method is developed for single-link flexible-joint robots with actuator failures.Specifically,on the basis of adaptive backstepping design,command filtered control is adopted to resolve the computational complexity problem and avoid the singularity problem in finite-time control.Next,an event-triggered mechanism with the relative threshold is adopted,so that the control signal can be intermittently transmitted to the actuator,which reduces the transmission burden of the signal channel.In addition,by estimating the bounds of relevant parameters in the fault model online,the proposed method can effectively compensate for the influence of actuator failures and quickly achieve satisfactory tracking control.Finally,the simulation results more intuitively verify the effectiveness of the proposed scheme.(2)The adaptive tracking control problem based on the event-triggered control is investigated for n-link flexible-joint robots with stochastic disturbances.First,the original system is transformed into the form of strict-feedback multi-input and multi-output stochastic nonlinear systems,and then a tracking controller is constructed by means of command filtered technology and backstepping design.Most importantly,due to the multi-input and multi-output structure characteristics,the method of designing the corresponding event-triggered mechanism for the control input component is adopted,and the proposed adaptive event-triggered control method can effectively improve the system performance.Finally,both theoretical analysis and simulation results demonstrate that the proposed scheme can obtain satisfactory tracking control and improve resource utilization at the same time.(3)A finite-time adaptive tracking control scheme is proposed for n-link flexible-joint robots with more general stochastic disturbances.Due to the consideration of more comprehensive disturbance factors,the transformed stochastic nonlinear system no longer has the lower-triangle structure.Compared with(2),in the process of combining command filtered technology and backstepping design to construct the controller,the fuzzy logic systems are employed to approximate the unknown uncertain items,and the difficulties caused by the nonstrict-feedback structure is resolved by means of the properties of the Gaussian membership function.In addition,the proposed scheme can quickly achieve good tracking performance of the controlled system by introducing a fractional-power term and a sign function term into the controller and the compensation signal,respectively.Finally,the simulation and comparison results demonstrate the effectiveness and superiority of this method.