An adaptive observer and its application to a pole placement controller

A discrete adaptive observer based on a weighted moving identification technique for an unknown linear SISO system is presented. It is an alternative to the scheme based on the exponentially weighted least squares method. This adaptive observer is then implemented in an adaptive pole placement controller which uses a state feedback control law for arbitrarily positioning all the poles of the overall system. In the weighted moving identification technique, two sets of weighting factors are used. One is for discarding the old data and the other is for adding new data. These weighting factors are so chosen to ensure that the proposed scheme exhibits a rapid convergence and tracking property for the linear time varying plant parameters. This scheme is demonstrated by computer simulations of a second-order plant