Model-based recursive Bayesian state estimation for single hydrophone passive sonar localization
- C. W. Jemmott
- PhD Dissertation, The Pennsylvania State…
We present a passive sonar tracking algorithm that utilizes variations in the amplitude of the signal received from a source in motion. The approach is model-based in that it employs an acoustic propagation computer program, and Bayesian in that prior knowledge is used to compute a posterior probability distribution function from which the target location is estimated. The algorithm is designed to be used on the continental shelf where the water is shallow (depth is less than 200 m). Low acoustic frequencies (< 1 kHz) are preferable because they suffer less attenuation due to absorption and because the lower spatial frequency is more robust to data-model mismatch. The method takes advantage of the multipath interference that occurs in shallow water, which causes a spatial interference pattern in the transmission loss (TL). We have successfully applied the algorithm to two shallow water sites. The current effort is to investigate how performance of the algorithm depends upon the quality or accuracy of the relevant parameters, including signal bandwidth, acoustic properties of the ocean bottom, and signal to noise ratio. Work supported by the Office of Naval Research Undersea Signal Processing.