Robust Autopilot Design Based on a Disturbance/Uncertainty Estimator

Abstract

Autopilot failures have caused fatal results for both military and civilian applications, leading to loss of life and property. Major contributing factors are disturbances (e.g., winds), uncertainties (e.g., unmodeled dynamics and parameter variations), and couplings (e.g., a bank maneuver affecting pitch attitude). In this brief, a novel disturbance/uncertainty/coupling estimator structure is introduced for identifying and canceling these effects as a whole. A controller scheme based on the $\mathcal {H}_\infty$ design is implemented with explicit bounds for robust stability and robust performance. Validity of the proposed approach is illustrated under high winds, strong couplings, and large parameter variations.