Modified model reference adaptive control: Design, analysis and performance bounds

From MaRDI portal





The performance of model reference adaptive control (MRAC) schemes in the absence of persistently exciting signals is generally characterized according to two criteria: boundedness of all closed loop signals and asymptotic convergence of the tracking error to zero or to a residual set. The authors of this paper address the issue of performance by applying two additional criteria: mean square tracking error and the \(L_\infty\) tracking error bound. The purpose for adding the latter criteria is to identify large transient oscillations as well as bursting at steady state in the presence of small modelling errors. The authors use their expanded performance criteria to motivate the design of a modified MRAC scheme that can have an arbitrarily improved nominal performance in the ideal case and in the presence of bounded input disturbances. The modified MRAC scheme can also provide arbitrarily improved zero-state transient performance and an arbitrary reduction in the size of bursts occurring at steady state. Real limitations to the performance of the authors' approach are the magnitude of unmodelled dynamics and sensor noise. For the entire collection see [Zbl 0864.00081].











This page was built for publication: Modified model reference adaptive control: Design, analysis and performance bounds

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4355816)