Time Delay Margin Analysis of Modified State Observer based Adaptive Controller


Adaptive control of systems with large uncertainties requires fast adaptation to ensure stability and good tracking performance. A drawback with using high adaptive gains, however, is that high gains could result in high frequency oscillations in the control signal. These oscillations in turn could excite the unmodeled dynamics of the plant and lead to system failure. An adaptive control methodology with a “modified state observer” using a generic observer structure is proposed to tackle the above problem. The use of modified state observer alleviates the oscillations in control signals by separating the design of estimation error dynamics and reference model dynamics. In this paper the effect of using modified state observer for adaptive control of input delay systems is analyzed. The local time delay margins are estimated using bounded linear stability analysis method and compared with that of standard model reference adaptive control. The relationships between estimation error dynamics, tracking error dynamics, adaptation gain and local time delay margin are investigated both analytically and numerically. Based on the analysis, an effective method is proposed for varying the adaptation rate online to ensure stable adaptation. Simulation studies were carried out for a simple scalar problem and also using a longitudinal pitch dynamic model of a Generic Transport aircraft. The results obtained indicate that the modified state observer based adaptive controllers are quite robust compared to MRAC.

Meeting Name

AIAA Guidance, Navigation, and Control (GNC) Conference (2013: Aug. 19-22, Boston, MA)


Mechanical and Aerospace Engineering

Document Type

Article - Conference proceedings

Document Version


File Type





© 2013 American Institute of Aeronautics and Astronautics (AIAA), All rights reserved.

Publication Date

22 Aug 2013