InverseOptimal NonlinearRecurrent High Order
Plaza LaLuna · 2005
Thispaperpresents thedesign ofanadaptive recurrent neural observer fornonlinear systems whichmodelisassumed tobeun- known.Theneural observer iscomposed ofaRecurrent HighOrder Neural Network whichbuilds anonline modeloftheunknownplant and alearning adaptation lawfortheneural network weights. Thislawisob- tained bytheLyapunov methodology. Thefeedback lawwhichguaran- teesstability oftheestimation error isproved tobeoptimal withrespect toawelldefined costfunctionaL