System identification and state reconstruction for autonomous navigation of an underwater vehicle in an acoustic net

Abstract

Closed loop control of an underwater vehicle in an acoustic position net requires an accurate hydrodynamic model of the vehicle. The model is essential to the control algorithm design process. Further, it is an integral part of the observer used to generate complete state estimates from the position measurements. An experimental apparatus and numerical analysis technique for vehicle system identification during the thruster induced hover limit cycle are described. Detailed comparisons to other techniques are made and extension of the technique to four degrees of freedom with coupling is discussed. A model of the Remotely Operated Vehicle Hylas is determined. The model determined by the system identification procedure is then used in the designs of a state estimator and controller for trajectory following by the vehicle. The algorithms are initially evaluated in a numerical simulation. Tests are made for stability, trajectory following performance, and accuracy of the state estimator under varying system and environmental conditions. Finally, the results of vehicle trials are presented. System stability and accurate trajectory following under the control of the algorithms are demonstrated using ROY Hylas. The high accuracy level of the simulation is also demonstrated by the trials and directions for continued research are discussed.