Underwater Model-Based Processing for SLAM and Environment Characterization

US military undersea assets are now required to operate submerged for long periods in shallow water environments where use of the global positioning system (GPS) for navigation information is not available. New techniques need to be developed to fuse novel forms of sensor information with vehicle inertial navigation systems (INS). This paper addresses the present US Navy investigation for using precision sonar sensing of the sea floor (bathymetry) and the acoustic environment to augment existing INS observations to enhance the performance of present systems. The fundamental algorithmic approach models the physics of all involved processes, and uses Kalman or particle filters as predictor-correctors. Benefits will be demonstrated via analyses and simulations. Simulation studies are presented which characterize the performance improvement achieved from various model based algorithms for underwater military platforms.