A Comparison Study of Several Adaptive Control Strategies for Resilient Flight Control

Abstract

In this paper we decribe our initial development and testing of the framework referred to as Algorithm Design & Validation for Adaptive Nonlinear Control Enhancement (ADVANCE). The key elements of ADVANCE are suitable performance metrics for adaptive control systems, and an automated tuning procedure using our in-house developed Automatic Tuner for non-Linear and Adaptive Systems (ATLAS). The paper describes a comparison study of the state-of-the-art adaptive flight control algorithms on two challenging testbeds. The development of the ADVANCE framework involved several steps: (i) Development of suitable performance comparison metrics for adaptive control systems; (ii) Development of a “plug-and-play” capability that enables rapid implementation and simulation testing of different advanced adaptive control algorithms; and (iii) Comprehensive simulation studies and performance comparison of the state-of-the-art adaptive control algorithms on a high-fidelity simulation of miniature tail-sitter UAV with significant nonlinearities and uncertainty, and a semi-nonlinear simulation of F/A-18 dynamics. Results presented in the paper demonstrate the feasibility and potential of the ADVANCE framework, and further development of the algorithms and testing procedures is expected to give rise to a set of recommendations and guidelines regarding the use, tuning and implementation of different adaptive flight control algorithms to different problems in flight control. This will also facilitate flight certification of adaptive flight control algorithms.