Improved Underwater Acoustic Voice Communication Based on Blind Separation

Abstract

Frogman underwater voice communication system is an indispensable equipment for underwater frogman combat. Among various methods of modulation, SSB voice communication system holds the advantages of high stability and better retention of speaker tone, however, with no capability of suppressing the underwater multichannel effect and being susceptible to noise. With the advancements in ocean development, there is a need for a voice communication system with both low voice distortion and robust transmit performance. In this paper, an acoustic voice communication technology based on blind separation is proposed, which uses complex cepstrum characteristic peak filter and minimum phase decomposition to calculate the amplitude and phase components of multichannel respectively. The performance of the improved voice communication system is verified by the simulated case and experiment data. In the experimental scenario, the average similarity coefficient between transmitted voice and processed voice can reach above 0.85, which reveals that the improved voice communication system can cope with the underwater multichannel effects.