Removal of coherent extremely low frequency (ELF) background noise by adaptive noise cancelation

The use of the Sequential Regression Algorithm (SER) to coherently remove background noise from an ELF sensor is presented. The SER algorithm is described for a multi-channel application in order to cancel coherent portions of reference sensors from a primary sensor. The algorithm adaptively accounts for differences between two parallel array platforms for the purpose of coherent subtraction. A section on likelihood ratio detector schemes for detecting narrowband signals is also presented. This work is in support of a submerged ELF sensor array project run by the Johns Hopkins University Applied Physics Lab