DETECTION OF A WEAK SINUSOID WITH DRIFTING PHASE IN NON-GAUSSIAN NOISE

The problem of detecting by an array of sensors a weak sinusoidal signal with drifting phase embedded in additive white non-Gaussian noise is addressed. The phase drift at each array element is modeled as a Brownian motion process. The locally optimum array detector is considered first. Then two suboptimum detection structures are examined: The first one is the locally optimum detector synthesized to detect a sinusoid with a random but constant phase; the second one is a version of the so-called mth-order noncoherent detector, which has been modified to account for the non-Gaussian behavior of the noise. The detection performances in terms of deflection are analytically evaluated and compared. Finally, to corroborate the results of the deflection analysis, the performances, in terms of detection probability for a fixed false-alarm rate, are assessed and compared by computer simulations. © 1994, Acoustical Society of America. All rights reserved.