Source localization at very long ranges in the presence of large-scale spatial variability

Abstract

Large-scale spatial variability in the ocean environment results in uncertain sound-speed variations on the order of 10–15 m/s, a potentially significant factor in many ocean acoustic applications. The effect of this variability on source localization performance at a range of one megameter is investigated. The optimum uncertain field processor (OUFP) [A. M. Richardson and L. W. Nolte, J. Acoust. Soc. Am. 89, 2280–2284 (1991)], in its narrow-band and wideband forms, is introduced as an alternate and superior processor to the traditional Bartlett processor for this application. The Bartlett processor, which assumes the environmental parameters to be the average values, essentially ignores the large-scale spatial fluctuations. Optimally incorporating a priori information about the environmental and source parameters, as does the OUFP, makes this processor robust to uncertainty in these parameters and provides an upper bound on localization performance. Measures such as a posteriori distributions of source location, the probability of correct localization, and the standard deviation of range estimation error reveal that the spatial variability has a significant degrading effect on localization performance, the magnitude of which is dependent on the source frequency. Results are discussed for 25- and 75-Hz sources as well as their coherent and random combinations. [Research supported by ONR (Ocean Acoustics).]