Key points are not available for this paper at this time.
The maximum-likelihood (ML) processor is presented for passively estimating range and bearing to an acoustic source. The source signal is observed for a finite-time duration at several sensors in the presence of uncorrelated noise. When the speed of sound in an isovelocity medium and the sensor positions are known, the ML estimator for position constrains the source to sensor delays to be focused into a point corresponding to a hypothesized source location. The variances of the range error and bearing error are presented for the optimum processor. It is shown that for bearing and range estimation, different sensor configurations are desirable. However, if the area of uncertainty is to be minimized, then the sensors should be divided into equal groups with one-third of the sensors in each group.
G. Clifford Carter (Sat,) studied this question.