Blind identification/deconvolution of single-event impulsive acoustic signals

Single-event impulsive signals abound in acoustic systems. They are characterized by a confinement of their energy in a short time interval, and they usually arise, intentional or unintentionally, from explosive acoustic sources. They can seldom be observed directly, as propagation and echoes superposition distort their original impulsive shape. This work addresses the recovery of the source signal in a multichannel framework, or the estimation of the impulse response function (IRF) when only one channel is available. The proposed blind deconvolution / identification scheme relies on a robust greatest common divisor (GCD) computation method.