Why Did Applications of MFP Fail, or Did We Not Understand How to Apply MFP?

Abstract

Matched Field Processing (MFP) was introduced in the 70’s by Bucker. Two experiments demonstrated localization at impressive ranges with vertical line arrays (VLA’s). MFP stimulated a plethora of articles in the 90’s and 00’s and a few experiments and lots of simulations. MFP suggested some enabling capabilities and was a great research topic; however, it generated a lot of “hype” for performance well beyond what anyone who really understood the acoustics and signal processing as well as potential navy concepts of operations (CONOPS) could believe. ONR and then ONT invested significant amounts of funding. In the 90’s ONT embraced the High Gain Initiative where MFP was supposed to be the answer to regain the lost dB for ASW surveillance. MFP did not perform well which led to too may questions for the Navy surveillance community to support MFP efforts. In retrospect the ap- plication of MFP in HGI was clearly premature. Many fundamental oceanographic, acoustic and signal processing issues were not understood and a lot more research was required then. This remains true. MFP is well suited for high SNR, well calibrated and low clutter applications, not the low SNR, uncertain and dense clutter applications typical of ASW surveil- lance. There were many experiments which clearly demonstrated that MFP could work. These included the NRL FRAM IV Arctic Experiment at 250 kms, the SAIC one at 1000 kms in the deep, eastern Pacific, the SPAWAR SWELLEX off San Diego and the Santa Barbara (SBCX), plus several others in both the US and Eu- rope. Nevertheless, the performance was not robust. In many others there were high sidelobes and incorrect localizations and the predictions of simulations simply were not demonstrated. There was never a clear answer why did MFP fail in these applications. This leads to the question dod we not understand how to apply MFP?