Neural Network-Based State Estimation Schemes
In this chapter, two neural network-based adaptive observers for a general model of MIMO nonlinear systems are proposed. The first proposed neural network is linear-in-parameter and the second one is nonlinear in its parameters which makes it applicable to many systems with arbitrary degrees of nonlinearity and complexity. The online weight-updating mechanism is a modified version of the backpropagation algorithm with a simple structure together with an e-modification term added to guarantee the robustness of the observer. The stability of the recurrent neural network observers are shown by Lyapunov’s direct method. Moreover, the strictly positive real (SPR) assumption imposed on the output error equation is relaxed.
KeywordsOutput Shaft Observer Model State Estimation Error Harmonic Drive Ultimate Boundedness
Unable to display preview. Download preview PDF.