Variable neural adaptive robust observer for uncertain systems

The design of variable neural adaptive robust observer is proposed for the state estimation of a class of uncertain systems. The proposed observer incorporates a variable-structure radial basis function (RBF) network to approximate unknown system dynamics. The RBF network can determine its structure on-line dynamically by adding or removing RBFs. The observer gain matrix is obtained by solving an optimization problem subject to linear matrix inequalities. The structure variation of the RBF network is taken into account in the stability analysis through the use of the piecewise quadratic Lyapunov function. The effectiveness of the proposed observer is illustrated with a simulation example.