Research On Optimal Output Feedback Control Based On Adaptive Dynamic Programming

In recent years,the optimal control problem has been one of the hotspots in the control field.The main idea is to find an optimal control strategy to optimize the value function composed of system state and control strategy.However,in the real industry,the state of the system is difficult to measure,so that the research on the optimal output feedback control becomes inevitable.In the study of optimal output feedback,the traditional solution only solves the problem of offline feedback and design an state observer to convert the output feedback problem into state feedback.To address the drawbacks of the traditional methods,this paper seeks a method to solve the problem online by means of the output state,and also rejects the scheme of designing the observer.This paper focuses on a mathematical transformation method combined with data-driven theory,so that the algebraic Riccati equation to be solved can be transformed into a problem solved by the output.In addition,in order to overcome the limitations of the first two methods,this paper combines the concept of state reconstruction to solve the output-feedback problem.The specific content of this paper is:Aiming at the optimal output feedback control problem of continuous-time linear systems,this paper first summarizes the optimal control problem and introduces the solution of state feedback control in the optimal control problem.At the same time,the relationship between optimal output feedback control and optimal state feedback control is introduced.The second chapter of this paper uses an offline solution method to solve the optimal output feedback control problem.Although the offline iterative method can solve the optimal controller value of the system according to the system structure,the offline iterative method can not meet the real-time requirement in industrial production.Therefore,in the third chapter of this paper,the online iterative method is proposed.The online iteration uses the output of the system to drive the ARE equation constructed according to the system.In this method,the gain matrix and system gain value of the system are defined as two asynchronous.The iteration value needs to give the gain value an initial value when iterating,in order to ensure that the iteration can continue.Although this solving method can satisfy the real-time problem of the system solution,the convergence speed of the controller is too slow because of its step-by-step iteration,which can not meet the requirements of the fast response of the controller in industrial production.Therefore,in the fourth chapter,this paper proposes an online adaptive scheme.This method uses adaptive to solve the optimal idea.When the algorithm is designed,the gain matrix and gain value of the control system are solved synchronously,so that online adaptation is a method of solving the problem,which can solve the controller more quickly.However,the system requires relatively harsh conditions and the solution accuracy is not fast enough.Therefore,the fifth chapter of this thesis puts forward the idea of state reconstruction.The state reconstruction mainly uses the output to reconstruct the state of the system,and then uses the refactored state to solve the optimal output feedback control problem.In order to illustrate the effectiveness of the proposed algorithm,the second,third,fourth,and fifth chapters of this thesis are simulated and analyzed.In the sixth chapter of this thesis,the three-degree-of-freedom helicopter experimental platform is used to analyze the third.The theory of the chapter was experimentally verified.