Adaptive matched field processing of a large array in a white noise environment

Two major problems occur in the adaptive processing of a large array. First, with present numerical algorithms, it is questionable whether a large array covariance matrix can be inverted with reliable accuracy. Second, the computer memory needed to store the covariance matrix is quadratically proportional to the size of the array. As the array size increases, the computer memory needed could easily exceed the hardware limitations of computer systems at present. To overcome these problems, analytical expressions for a subarray processing approach are derived for calculating the adaptive weight and the spectral output for a signal in a white noise environment. Sidelobe performance of different subarray configurations can be examined