Machine Learning based Dolphin Whistle Tranceiver for Bio-inspired Underwater Covert Communication

This paper proposes an underwater covert communication method using variety dolphin whistle patterns. The proposed method classifies dolphin whistles into G groups, and binary information is allocated to every consecutive different dolphin whistle. Received dolphin whistles are decoded by the random forest method and the transient probability of consecutive dolphin whistles. Computer simulation demonstrates that the BER performance of the proposed method is better than that of random forest method.