Application of statistical reduced isometry property to design of line arrays for compressive beamforming
Charles F. Gaumond, Geoffrey F. Edelmann · Proceedings of meetings on acoustics · 2013
The Statistical Reduced Isometry Property(StRIP) and Statistical Null Space Property (SNSP) are presented and reduced to numerical algorithms. These properties are used to predict the utility of a specific subsampled array for use in compressive sensing. Three examples of subsampling an equally spaced array are presented: random, Golomb and Wichmann. The Golomb array uses a Golomb ruler that has no repeated sensor element spacings. The Wichmann array includes at least one of every possible interval of sensor element spacings. The SNSP is shown to be insensitive to subsampling in the type of cases shown. The Golomb array is predicted to have superior performance to the Wichmann for comparable subsampling. The use of these two subsamplings for beamforming using at-sea data from the Five Octave Research Array (FORA) is shown. Research funded by the Office of Naval Research.