An optimal multisensor source location algorithm for passive sonar with moving source or sensors

An underwater acoustic source location algorithm, which is a globally optimal joint estimator of source position, course, and speed, in a maximum a posteriori probability (MAP) sense, is described. The algorithm accommodates multiple sensor arrays which need not be planar and which may themselves have motion. System motion is accommodated and restrictions on integration times are for practical purposes removed. Sensor noise may be nonstationary and is not required to be equal among the sensors of the array; the noise is modeled as white Gaussian, which can also apply to most colored noise cases by applying prewhitening. No practical constraints are placed on the spectrum of the acoustic signal, which need not be stationary. The algorithm is presented in the form of a focused beam former in combination with a filter correlator, but it may also be manipulated into filter squarer and correlator realizations. These general forms have also found use in suboptimal and asymptotically optimal estimators. Besides a direct implementation of the algorithm in system applications, it may be used as a performance reference in Monte Carlo tests to comparatively evaluate simplified or alternative algorithms, such as in the assessment of threshold performance.