Identification of non-linear dynamic systems using smooth variable structure filter

This paper proposes a model based approach for nonlinear dynamic system identification, which utilizes Smooth variable structure filter (SVSF) having multi fold advantages over other adaptive filters. The SVSF is based on sliding mode control concept and can be applied to both linear and nonlinear system. Further to increase the accuracy of nonlinear chaotic system identification the `memory element' γ is tuned iteratively in the measurement matrix of SVSF. So, a self tuning approach for updating memory element is proposed, which is inversely proportional to the sample number. Further SVSF is modified by introducing a nonlinear term with square innovation error in measurement matrix to achieve higher order sliding mode control. The effect and the improvement of this modification has been verified under different test conditions. Various simulations have been carried out to show the stability and improved robustness of the proposed algorithm over other conventional methods. The mean square error obtained using this approach is found to be significantly less than other conventional methods.