Multivehicles Cooperation: USV and AUV Cooperative Data Collection for Underwater Wireless Sensor Networks

To advance the development of maritime intelligent transportation systems (MITS), underwater wireless sensor networks (UWSNs), composed of numerous sensor nodes, have been widely deployed for underwater information perception. However, UWSNs face critical challenges in achieving cost-effective and timely data collection due to their large-scale deployment and stringent data timeliness requirements. To address this challenge, this paper proposes an efficient data collection scheme for UWSNs through the collaboration between uncrewed surface vehicles (USVs) and autonomous underwater vehicles (AUVs). Specifically, we first introduce a cooperative framework where AUVs select appropriate USVs to form USV-AUV clusters. Within each cluster, AUVs are responsible for sensing data collection, while USVs act as relay nodes, moving toward the destination (e.g., data center). We then devise an evolutionary game-theoretic cluster forming mechanism, deriving evolutionarily stable strategies (ESS) through replicator dynamics analysis, which guarantees a provable Nash equilibrium. Next, we present a hierarchical optimization method that models the interaction between UWSNs and the cluster as a two-agent Markov decision process, where a dual-agent Q-learning algorithm is designed to jointly optimize the decisions of both entities. Finally, extensive simulations demonstrate that the proposed scheme outperforms conventional methods in improving the efficiency of sensing data collection for UWSNs.