Neuroadaptive PI Control For Nonlinear Systems With Input And State Constraints

In recent years,with the further study for the actual engineering control systems,strict requirements has been exerted to the states of the systems as taken performance,security and physical saturation into consideration.If undesirable state values occur in the system operation,the operation of the whole system may be unpredictable.Therefore,more and more scholars begin to study the state constraints problem for the control systems.Considering that the actual engineering system is mostly non-linear and has many uncertainties(such as unknown parameters,measurement error and external disturbance,etc.)as well as inevitable abnormal actuation characteristics(including input dead-zone,saturation and actuation faults,etc.),in this work,we propose a neuro-adaptive PI control method for uncertain nonlinear systems with full state constraints,which are simplicity and intuitiveness in both structure and concept.For a class of uncertain nonlinear systems with input dead-zone and saturation,a neuro-adaptive PI control is proposed.In order to deal with the modeling uncertainties and the impact of the abnormal actuation characteristics,neural networks are utilized at each step of the backstepping design.By using barrier Lyapunov function,together with the concept of virtual parameter,a neuro-adaptive control scheme is developed so that both tracking stability and full-state constraints are ensured all the time.Finally,the effectiveness of the control method is verified by simulation,and it can be seen that the proposed control strategy has better robustness and steady-state performance than traditional PI control.For a class of uncertain nonlinear systems with input dead-zone,saturation and actuation failures,a neuro-adaptive PI control with guaranteed transient and steady-state performance is proposed.First,a performance transformation is carried out so that the problem of ensuring asymmetric performance constraint can be transformed into the stabilization problem of the conversion error.Second,the neural networks are utilized at each step of the backstepping design in order to deal with the modeling uncertainties and unknown functions.Third,by using barrier Lyapunov function,together with the concept of virtual parameter,a neuro-adaptive control scheme is developed to ensure stable tracking control and guaranteed transient and steady-state performance of the system,meanwhile,all the internal signals are continuous and bounded.Besides,the full-state constraints are satisfied all the time.By comparing with the numerical simulation results of traditional PI control,the proposed one shows better robustness,transient and steady-state performance.