ReFOCUSD: a technique for real-time enhancement of low-frequency SONAR images in dynamic situations

Arion Systems, Inc. has developed an optimal adaptive array processing technique, called ReFOCUSD, capable of producing high-resolution SONAR images in applications of a transient or dynamic nature. The technique recursively optimizes the array weighting functions based upon successive estimates of the noise field strength distribution. The initial estimate of the field strength can be an independent measurement, an analytical or empirical model or a simple assumption. The definition of the weighting functions has a closed form relationship to the source distribution and is not based upon direct measurements of the sensor signals. As such, the individual sensor signals can be relatively noisy and the exact sensor locations can be uncertain. The high resolution images are produced real time for nonuniform array geometries with many elements and in dynamic situations involving targets in motion, host platform motion or deformable arrays. This paper introduces the ReFOCUSD algorithm and discusses its capabilities relative to other optimal adaptive methods. Additionally, the capability of ReFOCUSD to produce high-resolution images in time-bandwidth limited situations is demonstrated in two examples. The first example involves a numerical simulation of an existing passive SONAR system currently in use by the US Navy. The second example gives experimental results obtained by processing in-air acoustic data associated with vortex-shedding noise from a blunt trailing-edge airfoil provided by the University of Notre Dame.