Tensor Product Model-Based Gain Scheduling of a Missile Autopilot

Abstract

In this paper, a tensor product model-based gain scheduling technique is utilized to design a pitch-axis autopilot for an air-to-air missile. Firstly, a recent technique for tensor product (TP) model transformation provides a convenient way to transform missile pitch-axis linear-parameter-varying (LPV) model into a convex parameter-varying weighted combination of linear-time-invariant (LTI) systems. This polytopic model is beneficial for convex hull manipulation, so that a large number of linear matrix inequality (LMI) optimization techniques can be applied controller development. This paper presents an alternative LMI technique based on a TP polytopic model, which can optimize the H_∞ performance of a closed-loop system with LMI pole constraints. Then, the proposed approach is applied in the design of a simple four-loop gain-scheduled autopilot. Final simulation results indicate the missile autopilot presented has good overall performance and strong robustness, which validates effectiveness of the proposed TP model-based method.