查看文章

In this paper, regarding the observer form of the well-known Youla parameterization, the controller design and optimization are exhibited with an integrated residual access. To better reveal this philosophy, the feedback control loop is interpreted on the basis of the observer-based residual generator. The next main attention is drawn to the generation of residuals, the design of a deadbeat controller for system stabilization both in the data-driven environment, and later the optimal adaptive realization of a dynamic system that translates residuals into compensatory control inputs to meet certain performance specifications. Towards these goals, numerical algorithms are summarized, and for the issues of controller optimization, the reinforcement learning algorithm is introduced using only measured input-output and residual signals. In addition, the effectiveness of developed schemes for industrial applications is also …