Preliminary experiments in the adaptive identification of dynamically positioned underwater robotic vehicles

This paper presents a stable online adaptive identification technique for the identification of finite-dimensional dynamical models of dynamically positioned underwater robotic vehicles. A direct comparison of this method to a conventional, off-line, least-squares method is presented. Based on experimental data obtained using the JHU remotely operated vehicle, both methods are employed to develop decoupled, single degree of freedom dynamical plant models. Performance of the resulting dynamical models is compared to the logged experimental motion of the actual vehicle.