Aircraft Vision-Based Landing Using Robust Extended Command Governors

Abstract

We propose a novel vision-based constrained control solution to execute aircraft landing on an unknown runway. The method is based on an extended command governor which is tailored for a specific class of constrained linear systems with a single uncertain modeling parameter. The computational burden is reduced by exploiting a robust strongly returnable set in place of a robust invariant set and computing this set through the use of a binomial expansion. Numerical results illustrate the ability to ‘safely’ align the aircraft with the runway while satisfying the specified terminal constraints to execute the flare. Here the terminal constraints are efficiently handled by transforming them into time-varying constraints.