Improvement Research Of Model-free Control Algorithm Based On Data And Model Driven

Model-free control method based on the data-driven control theory can control acomplex system, which is independent of systemic mathematics model and only needinput-output data. The method has some of innovations in theory. Therefore, the developmentof data-driven control theory has important theoretical and practical significance. In this paper,to overcome the faultiness of universal model theory and control law without model, we dothe research and improvement work. The details are as follows:(1) Deeply researched model-free control algorithm and the model-free control improvedalgorithm based on the theory of multi-innovation proposed by our team, the paper gives itsconvergence proof by approximation algorithm.(2) For the improvement of model-free control algorithm based on the theory ofmulti-innovation, the work is mainly focused on:Utilizing the principle of error feedback, the model-free control improved algorithmbased on multi-innovation with error feedback is proposed. It not only use input and outputdata of the system effectively, but also use error to modify universal model improving thecontrol property. It can improve the convergence performance and may effectively solve theproblem of system overshoot.According to on-line optimization for model free adaptive control algorithm based on thetheory of multi-innovation. This paper use drosophila optimization algorithm to realizeon-line optimization of controller. The simulation results show this method is effectively.(3) Study the PID neural network identification and control theory, PID neural networkalgorithm based multi-innovation theory is proposed, and we theoretically prove theconvergence of the improved method. Finally, simulation experiments demonstrate that themethod is effective.(4) Combining the detail project, the paper deeply analyzes the Stewart model, open control system and related theory for the problem of the attitude adjustment of six degrees offreedom adjustable pier which has characters of time delay, complex and nonlinear. Theimproved algorithm proposed by this paper is used to control the system of Stewart, practicalapplications demonstrate that the improved method can effectively solve the six degrees offreedom adjustable platform control problems.