Anti-jamming Communication in UWSNs with Dynamic Spectrum Waterfall: A Double-DQN Approach

A dynamic spectrum waterfall-based Double- DQN quick convergence anti-jamming algorithm (QCAA) is proposed for the problem of anti-malicious jamming in Underwater Wireless Sensor Networks (UWSNs). Without a priori knowledge of the jamming strategy of the jammer and the complex underwater communication environment, in order to improve the convergence speed of the anti-jamming decision- making process for underwater communications, the QCAA algorithm is used in conjunction with Long Short-Term Memory (LSTM) to solve the optimal anti-jamming strategy. Simulation results show that the proposed scheme can effectively improve the normalized throughput of UWSNs under four malicious jamming scenarios compared with the conventional DQN-based channel selection scheme and the random frequency hopping anti-jamming scheme.