The Robust Control Method Research of a Combined Controller Design with QFT and PID Based on BP Neural Network

摘要

For the plant with uncertainty and high non-linearity, this paper puts forward a new combined controller design method with quantitative feedback theory and PID controller based on BP neural network. First, a method combined the BP neural network and the tradition PID control is developed, an initial controller is given. Then when the controller quality can not be improved effectively by tuning close-loop weighting function ，QFT is used to shape an open-loop transfer function according to the performance index bounds. The shaping has transparency in Nichols Chart and tailor characteristic on special frequency.The intermediate frequency characteristic of system is controlled by turning ξ and n ω .The high frequency performance is compensated in the QFT framework. The simulation result reveals that this method can generalize to design the control system in same uncertainty and high non-linearity system.