| This study introduces a method for using a single underwater sensor and rough knowledge of a realistic shallow water environment to track a moving, multi-frequency, acoustic source. The motion of the source creates a synthetic aperture that is adaptively matched against an acoustic propagation model in a manner similar to matched field processing. However, in this case, many more unknowns must simultaneously be determined. Results are presented for experimental data without prior knowledge of source speed, heading, level, and initial position. A new parallel adaptive method manages these simultaneous unknowns. Perturbations in the sound speed based on empirical orthogonal functions are used to extend the approach to operating when there is also uncertainty in this parameter. The performance of the tracking process in various signal-to-noise cases is considered. |
