Hybrid state estimators for the control of remotely operated underwater vehicles

Abstract

This paper explores the use of 'hybrid' state estimators to increase the accuracy and flexibility of acoustic position measurement systems used in the control of underwater vehicles. Two different approaches to extend the range of acoustic position measurement systems are explored. The first approach is the use of an inexpensive Strapdown Inertial Measurement System (SIMS) to augment the acoustic with inertial position information. This approach is based on the experience gained using an attitude and inertial measurement package fielded on the JASON JUNIOR Remotely Operated Vehicle (ROV). The second approach is the use of a mobile, platform-mounted, acoustic net in conjunction with a platform tracking system. This second investigation used the JASON ROV as the basis for the simulation work. As motivation, some of the theoretical and practical difficulties encountered when range is extended using an unaugmented system are explored. Simulation results are used to demonstrate the effects of these difficulties on position estimation accuracy and on the performance in closed loop control of the vehicle. Using measured sensor noise characteristics, a hybrid Kalman filter is developed for each approach to take the greatest advantage of the available information. Formulation of the Kalman filter is different for each case. In the second case, the geographic position of the ROV is the sum of the acoustic net's geographic position, measured at a different interval by an RF positioning system, and the position of the ROV relative to the net, as measured acoustically. Closed loop vehicle performance evaluations are made for representative noise levels and update rates with and without the augmentation discussed in the first approach. Finally, conclusions are drawn about the benefits and applications of the hybrid Kalman filter to the control of Remotely Operated Vehicles.