Robert Zeskind
The MITRE Corp., 7525 Colshire Dr., McLean, VA 22102-3481
Mark Owen
PRESEARCH, Inc., Fairfax, VA 22031
Matched field processing (MFP) is based on some type of comparison of actual received signal across the elements of an array to predicted array element response based on complicated propagation models for postulated source locations. It is well known that these techniques are very sensitive to model mismatch. Investigators have proposed modified MFP techniques of varying complexity to be more robust to mismatch. A simplified approach to MFP in a correlated multipath environment is to estimate the parameters of the path arrivals at the array based on an assumed wavefront model for each path. The parameters to be estimated are arrival angles, relative received level for each path, and the correlation coefficient between pair of paths. For spherical wavefronts, the distance traveled along each path is also estimated. It is assumed that there are only a few dominant paths and their number is known. The parameter estimation problem is posed as a well-known optimization problem. The solution to the optimization problem not only acts as a detector, but provides the set of estimated parameters to be used in a raytrace propagation model to back propagate the multipath to localize the detected target. An example is presented to illustrate the method.