Advances in multibeam sonar mapping and data visualization have increasingly brought to light the subtle integration errors remaining in bathymetric datasets. Currently, traditional field calibration procedures (the patch test) just account for static orientation bias and sonar to position latency. However, this ignores the generally more subtle integration problems that generate time varying depth errors.

Such dynamic integration errors are the result of an unknown offset in one or more of orientation, space, sound speed or time between the sonar and ancillary sensors. Such errors are systematic, and thus should be predictable, based on their relationship between the input data and integrated output. A first attempt at addressing this problem utilized correlations between motion and ping-averaged residuals [1]. The known limitations of that approach, however, included only being able to estimate the dominant signal, imperfectly accounting for irregular sounding distribution and it only working in shallow water.

This paper presents a new and improved means of utilizing the dynamics of the integration error signatures which can address multiple issues simultaneously, better account for along-track sounding distribution, and is not restricted to the shallow water geometry. Through considering soundings from extended sections of a single swath, the motion-driven signatures of multiple integration errors may be simultaneously identified. Such an approach provides a means of underway system optimization using nothing more than the bathymetry of typical seafloors acquired during transit. Successful estimation, however, imposes conditions of significant vessel motion, and smooth, gently rolling bathymetry. Initial results of the new algorithm are presented using data generated from a simulator (with known inputs and integration errors) to test the efficacy of the method.