Cost Detectability and Stability of Multiple Model-Based Adaptive Controllers using Data-Driven Control Theory


Adaptive systems aid in achieving desired performance even in the presence of uncertainties. However, their use in safety critical systems like aircrafts is limited due to the non-availability of tools to comprehensively evaluate in advance, the performance and robustness of the designed adaptive controller. To overcome the above difficulties, this paper proposes an online data-driven verification and tuning method. In this approach, multiple adaptive control schemes run in parallel and the best scheme is selected at each step based on collected data. Optimal switching between different adaptive schemes is achieved through a data-driven control theory concept called 'Unfalsified adaptive control'. This collaborative adaptive control scheme is expected to provide superior performance as compared to a single adaptive control scheme. Furthermore, a class of data-driven cost functions that are cost detectable is used to detect instability. This in turn improves the robustness of the switching control, for candidate adaptive schemes. Simulation results demonstrate the efficacy of the proposed algorithm.

Meeting Name

2017 AIAA Guidance, Navigation, and Control Conference, AIAA SciTech Forum, MGNC 2017 (2017: Jan. 9-13, Grapevine, TX)


Mechanical and Aerospace Engineering

Keywords and Phrases

Active safety systems; Control theory; Controllers; Cost functions; Costs; Robustness (control systems); Safety engineering; Adaptive control schemes; Adaptive controllers; Data-driven control; Multiple-modeling; Optimal switching; Safety critical systems; Switching Control; Unfalsified adaptive control; Adaptive control systems

International Standard Book Number (ISBN)


Document Type

Article - Conference proceedings

Document Version


File Type





© 2017 American Institute of Aeronautics and Astronautics (AIAA), All rights reserved.

Publication Date

01 Jan 2017