A U-model based controller design for non-minimum phase systems: Application to Boeing 747 altitude-hold autopilot

This paper presents a novel approach to design non-minimum phase (NMP) dynamic control systems, which consists of two stages of design. Specify a desired minimum phase (MP) plant model, with desired zeros and the same orders as original plants and then determine a hauling polynomial (by resolving the difference between the specified MP zeros and underlying NMP zeros) to convert a NMP control system design into MP formality. The second stage is to design an invariant closed loop system, then find the controller output by resolving an inverse dynamic in form of U-model root solving. To demonstrate the design procedure and potential for applications, take Boeing 747 altitude-hold autopilot for bench test. For comparison, Linear Quadratic Gaussian (LQG) design is selected which is a very classical approach. Simulation results illustrate the effectiveness and efficiency of the proposed control scheme.