Adaptive Robust Servo LQR Control for Aircraft Under a Wide Range of Icing Conditions

Abstract

Icing can significantly affect aircraft stability and control derivatives, compromising flight safety and performance. Aircraft may encounter a wide range of icing conditions, from moderate to severe. To address this, autopilots must handle aircraft recovery in icing conditions. In this study, we propose an Adaptive Robust Servo Linear Quadratic Regulator (ARS-LQR) approach to simultaneously control the aircraft's altitude and forward velocity under a wide range of icing conditions. The ARS-LQR approach, which obtains linear models by linearizing around the local trajectory, may seem classical, but it introduces a practical element by incorporating an exponential multiplier into the cost function. This element ensures local stability with a relatively large stability margin, even in severe icing conditions, making it particularly valuable for aircraft control engineers dealing with certification requirements. This theoretical support distinguishes it from many recent techniques, including AI-based ones, lacking such backing. A comprehensive comparison study validates the theoretical advantage of ARS-LQR over Adaptive Neural Nonlinear Dynamic Inversion (AN-NDI) in tracking altitude and forward velocity references in severe icing conditions. Existing control techniques do not address the full spectrum of aircraft icing conditions. Computer simulations emphasizing longitudinal motion provide an efficient and stable solution for severe icing challenges.