The multichannel maximum-likelihood (MCML) method: a new approach for infrasound detection and wave parameter estimation

SUMMARY We are presenting a new and novel approach to the detection parameter estimation of infrasonic signals. Our is based on likelihood function derived from multisensor stochastic model expressed in different frequency channels. Using function, we determine, for problem, generalized ratio (GLR) slowness vector, maximum (MLE). establish asymptotic results (i) GLR under null hypothesis leading computation corresponding p-value (ii) MLE by focusing two wave parameters: backazimuth horizontal trace velocity. The multichannel maximum-likelihood (MCML) method implemented time–frequency domain order avoid presence interfering Extensive simulations with synthetic signals show that MCML outperforms state-of-the-art correlation detector algorithms like progressive terms probability false alarm rate poor signal-to-noise scenarios. also illustrate use real data International Monitoring System how improved performances this lead refined analysis events accordance expert knowledge.