GPU-Based Parallel Implementation of Fisher Detector for Infrasonic Spectrum Sensing Systems

Abstract

Fisher statistics-based signal detection is a widely-used powerful multi-sensor infrasound spectrum sensing method. However, this method requires the repeated computation of test statistics for each element of a grid of slowness vectors, which imposes a high computational complexity and leads directly to a raised processing time. Since conventional systems often have very stringent speed requirements for real-time surveillance applications, this disadvantage leads a limited application of Fisher detectors (FDs) for several infrasonic sensing purposes. In this paper, we propose a strategy for implementation of FD with reduced time-consumption. This strategy is based on the fact that the detection process for slowness-grid elements can be performed in a parallel manner using powerful graphics processing units, in contrast with conventional FDs. The results of simulations show that our strategy promises very significant time-savings compared with conventional FDs, which enables it as a good candidate for real-time applications.