Probability Inference Approach for Stochastic Robust Flight Control Design

Abstract

This paper proposes to apply probabilistic classification technique to stochastic robust flight control design. Stochastic robust control design is an approach to optimize a flight controller based on Monte Carlo Evaluation. The region in the parameter space satisfying design requirements with high probability is estimated using Bayesian inference and is used to search the optimal design parameters. The approach has the potential to reduce computational load in stochastic optimization. It is applied to a simple example problem of flight control design and the result is promising.