Optimal Control Framework for Estimating Autopilot Safety Margins

Abstract

This paper presents an optimal control framework to determine a collection of open-loop command signals that mathematically guarantees operation of an aircraft within certain prescribed state constraints. The framework is specifically applied to estimate margins for the reference command inputs of aircraft autopilot systems so that safe operation within a given flight envelope can be assured under appropriate control action. Flight envelope excursions are generally considered as precursors to loss-of-control incidents; hence, these margins contain safety-critical information that can help improve the situational awareness onboard the aircraft. In off-nominal conditions, the computed safety margins provide indications of a degraded aircraft with reduced flying and handling qualities. These indications appear in the form of increasingly more strict limits on the autopilot reference command input. The entire framework is illustrated on an example problem involving a pitch dynamics model with state constraints on the pitch attitude. Simulations are conducted wherein margins are computed for the reference pitch command of the pitch hold system, while the aircraft enters an off-nominal condition with severely degraded system dynamics and reduced elevator effectiveness.