Abstract:

Recent studies with simulations and real data have shown that multiple frequency information can lead to increased robustness in matched-field estimation. This observation generates the question of how this information should be combined for best results. In this work an estimator is proposed that is suitable for matched-field inversion involving multitonal sources. The processor operates on the basis of an interfrequency spatially coherent framework. Designed for passive problems, it does not require knowledge of the source spectrum, an advantage over other coherent schemes. Moreover, the processor has the additional advantage of a simple implementation. Using synthetic data, the new coherent processor was compared to the conventional Bartlett processor with incoherent averaging across frequencies. The results indicate that the new processor is, in general, preferable to the conventional Bartlett processor. Interestingly, the superiority of the coherent processor is most pronounced when the receiving array undersamples the propagation medium, making the processor attractive for its cost effectiveness. [Work supported by ONR.]