Improved data driven strategy for aircraft controller design

This paper proposes the improved data driven strategy for aircraft control system, including three aspects, i.e. data driven identification, data driven control and data driven validation. Firstly, after reviewed the classical aircraft flight control system as one closed loop system with the unknown aircraft model and controller, within the framework of perfect tracking, controller is yielded through solving one optimization problem. To give one explicit form of cost function, data driven identification is applied to construct one auxiliary model, replacing the unknown aircraft model. Secondly, to avoid the modeling process, data driven control is applied to design one parameterized control, whose controller parameters are tuned by gradient algorithm. Thirdly, to testify whether the designed controller work well, data driven validation is applied to the correlation based validation method. Finally, the detailed derivations about our proposed data driven strategy are given from the point of theoretic aspect, and its application in aircraft flight system is also given. Generally, this paper combines our new contributions from the theory of data driven strategy and its practical application.