The Controller Design And Analysis Based On Characteristic Model

Characteristic modeling and the all-coefficient adaptive control method based on characteristic model, presented by academician Wu Hongxin after experiencing in prac-tical engineering application for years, are new practical and systematic control theory and method. Characteristic modeling takes plant dynamic characteristics, environmen-tal characteristics and control performance requirements into consideration to get a sim-ple and low order model. It avoids the difficulty in designing controller when establish-ing accurate dynamic model. Due to the simple design and operation in engineering, the novel control method has been greatly applied in practice. In this paper, we will talk about the controller design based on a second-order differential characteristics model and the related work. This paper mainly involves the following three issues:(1) The design of a pole placement adaptive controller and stability analysis of the corresponding closed loop system are discussed for a class of SISO systems based on the characteristic model method. Firstly, constructing a second-order slow time-varying linear system as the characteristic model of original system by a compress map-ping function; Secondly, based on characteristic model, extending the golden section adaptive controller form and proposing the pole placement adaptive controller; Finally, taking the error between characteristic model and original system’s accurate discrete-model into consideration, we give the sufficient conditions of stability for the closed loop system by using the related conclusions of sample system and perturbation analy-sis.(2) In view of some potential conditions in the characteristic modeling process, We study the controller design based on characteristic model with taking the constraint into account. We translate the conditions asked when constructing characteristic model from the original continuous system into constrains by use of its related state or input at first. Further, we re-description the constraints in form of characteristic model’s; then we present a model predictive control scheme based on the identification problem of characteristic model. In order to solve the problem that the golden section adaptive controller may fails to stabilize the nonlinear system under a fixed sampling period, we propose a phased control method based on characteristic model.Finally, simulations show the effectiveness of the above two control methods.(3) The extended application of characteristic model theory; Firstly, according to the linear system identification problem in the presence of randomly missing mea-surements, we study a Kalman filter-based identification for system by taking the all-coefficient theory into account. By doing so we show that the accuracy of the estimation is improved, both theoretically and numerically; Secondly, for the linear system with a disturbance is imposed on the control input channel, we present a golden-section adap-tive control scheme based on the characteristic model with a disturbance observer. A simulation example is given to demonstrate the effectiveness of the method.