Application of Robust State and Parameter Estimation to a Supercavitating Torpedo Model

This document describes a robustness enhancing modification to an existing joint state and parameter estimation technique as well as an intuitive and practical method to enhance on-line parameter identification by reducing correlations amongst the identification inputs. Results from the application of these ideas to a supercavitating torpedo model are included. The background of the robust joint estimation enabling technologies as well as their inte- gration is described first followed by an example demonstrating the enhanced robustness to noise and insucient excitation. Subsequently, the robust joint estimation approach is employed to provide state and parameter estimates to a nonlinear controller which gener- ates eector commands for the torpedo model. Next, the correlation reduction approach to enhance on-line parameter identification is introduced and analyzed followed by simulation results to demonstrate the ecacy.