The Mixing Zephyr Pages

Mixing Zephyr Program
Navigation Issues

High precision navigation of ALVIN is an essential component of this dive program, and will be necessary within the context of RIDGE observatory measurements in the coming decade for both ALVIN, ROVs, and other towed and moored platforms. Navigation within a long baseline acoustic transponder network is generally used for this purpose. In 1988 we initially developed the capability to use sub-transponder-sub travel times from pairs of transponders to obtain ALVIN positions that were precise within ~1 meter. While this represented a major advance in our ability to map and sample at fine scales, there were two important limitations. The accuracy of the fixes were limited by the quality of the survey which could be performed from the surface ship--while the repeatibility of positions was excellent there was no assurance that the grid was rectilinear. Indeed that offsets would exist between the positions calculated from different pairs of transponders demonstrates that the transponder positions were not known with sufficient accuracy. When re-establishing navigation nets during subsequent cruises there were competing demands to have the new navigation in registration with existing maps of the vent fields, versus working with the improved knowledge of transponder locations as surveyed from the support vessel brought about by the evolving technology used to establish the ship position (LORAN to GPS to p-coded GPS).

A second issue relates to the analysis of the survey data. The computational techniques used by the ALVIN group to survey transponders has undergone a series of changes. The present technique is to work transponder-by-transponder, minimizing the error in range as the ship locates itself at different positions around the transponder. We found (during the Coax response cruise in 1993), that this procedure was unacceptable given the offsets inherent in conventional GPS positions, leading to errors in transponder locations of order 25 meters which would translate in offsets between positions computed from different baselines of order 100 meters. In principle, these problems would be solved with use of p-coded GPS navigation for the surface ship which would be available for our program. However one still would have uncertainties brought about by errors in the ray-bending calculations necessary to convert travel times to range.

In the summer of 1994, we began discussions with Bruce Howe of the UW Applied Physics Laboratory on the applicability of software he had developed for performing acoustic tomography experiments for the purpose of navigating the submersible (his application was correcting for mooring motion during tomographic experiments). A desireable attribute of his software was that it would allow us to use data from both the ship and the submersible to compute transponder locations. With the ship data, we would be able to locate the net in absolute coordinates, while the submersible data would allow the most precise relative positioning of the transponders because of having minimized the errors inherent in ray-bending computations. In this section we summarize the results of our experiences with this new approach.

MORP Home | Mixing Zephyr Home | Search the Mixing Zephyr Pages

The Mixing Zephyr Pages
Maintained by Russ McDuff (
Copyright (©) 1995,1996 University of Washington; Copyright Notice
Content Last Modified 8/4/96
Page Last Built 8/4/96