A trust management-based secure routing protocol with AUV-aided path repairing for Underwater Acoustic Sensor Networks

Underwater Acoustic Sensor Networks (UASNs) are widely applied in marine based applications, but traditional security mechanisms are not feasible due to the underwater sensor nodes' limited constraints and capabilities. Moreover, few existing routing protocols for underwater sensor networks considered underwater transmision limited constraints. In this paper, we propose T-SAPR, a trust management-based secure routing protocol with AUV-aided path repairing that addresses these challenges. T-SAPR uses an attention-based long short-term memory (LSTM) to build a trust model that considers node trust, communication trust, and environment surrounding's trust. T-SAPR evaluates trust evidence to measure the trust level of the sensor nodes in the network and uses energy, packet delivery ratio, and latency jointly to seek optimal routing policies in the reward function of Q-Learning. Additionally, we propose a repairing mechanism with the assistance of AUVs to improve the reliability of UASNs in cases of multiple sensor node failures or the detection of malicious nodes. Evaluation results show that T-SAPR identifies malicious nodes by detecting their anomalous behavior while improving the packet delivery ratio and energy efficiency by 13.3% and 48.1% as compared to existing algorithms such as GEDAR, TADR-EAODV, and AFSA-ACOA-SCP.