Model-Based Positioning

Date/Time:

13 Mar 2012
14:00-14:30

Location:

South Galleries

Seminar Details
Generally, underwater positioning and navigation is a rapidly growing field of research since information about the current position of an underwater object like a diving robot or a marine mammal is very useful for a wide variety of applications. Examples comprise ocean science and exploration, underwater archaeology and salvage, as well as offshore mineral mining and extraction. Presently, most acoustic underwater positioning systems use the straight path assumption to determine the distances between transmitter and receiver, e.g. between an underwater vehicle and a surface buoy. This means that distances are calculated from the estimated travel time of an acoustic signal and the speed of sound, which is often assumed to be constant. However, sound waves typically do not follow a straight line through the water due to refraction. Therefore, the estimated distances may differ significantly from the true distances, i.e., the straight path assumption is a major error source concerning acoustic underwater positioning. In this paper, a model-based positioning approach is proposed, which takes into account the acoustic ray bending caused by refraction. Based on the BELLHOP acoustic beam/ray tracing model the estimated transmitter-receiver distance is corrected iteratively in order to deliver more accurate location estimates. A related approach is applied in NASNet (Nautronix Acoustic Subsea Network), which integrates refraction into long baseline array planning based on ray tracing. However, ray bending is not considered for positioning purposes in NASNet. The proposed model-based positioning approach is able to improve distance estimates for shallow- and deep-water environments significantly, if perfect environmental data is available at the beam tracer. Compared to the straight path assumption the distance error could be improved by a factor of 7.4 on average. Typically, in deep-water environments the modelling error is virtually zero, as well as in most shallow-water environments under investigation.