Anti-Jamming Underwater Transmission With Mobility and Learning

Abstract

In this letter, we present an anti-jamming underwater transmission framework that applies reinforcement learning to control the transmit power and uses the transducer mobility to address jamming in underwater acoustic networks. The deep $Q$ -networks-based transmission scheme can achieve the optimal power and node mobility control without knowing the jamming model and the underwater channel model in the dynamic game. Experiments performed with transducers in a non-anechoic pool show that our proposed scheme can reduce the bit error rate of the underwater transmission against reactive jamming compared with the $Q$ -learning based scheme.