Automatic weighting filter tuning for robust flight control law design

This article presents a practical framework for the co-design of weighting filter parameters and controller gains in the context of multi-objective/multi-model tuning of flight control laws. A non-smooth optimization-based procedure is proposed aiming to reduce the complexity of the control design process, evolving from a rather classical manual filter selection all the way to a fully automatic one. A normal acceleration Control Augmentation System (CAS), designed by making use of non-smooth optimization software, for the linearized dynamics of an agile autonomous aerospace vehicle is used as an illustrative example. Several performance indexes are compared in the context of the developed framework as well as against more classical tuning methods such as pole placement, demonstrating that the proposed approach may lead to improved results and potentially benefit the flight control engineer via an easier handling of conflicting design requirements, model uncertainty and application-tailored control structures.