| Zhang Yuzhi (张育芝)* **,Jia Menglei* **,Han Xiang**.[J].高技术通讯(英文),2026,32(3):231~239 |
|
| DRL-based time-domain interference alignment MAC protocol for underwater acoustic networks |
| |
| DOI：10. 3772 / j. issn. 1006-6748. 2026. 03. 002 |
| 中文关键词: |
| 英文关键词: deep reinforcement learning, underwater acoustic network, medium access control protocol, interference alignment |
| 基金项目: |
| Author Name | Affiliation | | Zhang Yuzhi (张育芝)* ** | (*College of Communication and Information Engineering, Xi’an University of Science and Technology, Xi’an 710054, P. R. China)
(**Xi’an Key Laboratory of Network Convergence Communication, Xi’an University of Science and Technology, Xi’an 710054, P. R. China) | | Jia Menglei* ** | | | Han Xiang** | |
|
| Hits: 61 |
| Download times: 79 |
| 中文摘要: |
| |
| 英文摘要: |
| To improve network throughput in clustered underwater acoustic networks, this paper proposes a
multi-agent deep reinforcement learning-based time-domain interference alignment medium access
control (MAC) protocol (DRL-TDIA-MAC). The theoretical throughput upper bound is derived by
utilizing the underwater long propagation delays to achieve interference alignment across multiple
clusters. In the proposed DRL-TDIA-MAC, the interference alignment information is used to refine
the state information and enhance the reward mechanism, thus guiding the DRL agent to converge to
the theoretical optimum throughput. In addition, a reward-stratified experience replay mechanism is
introduced to enhance sample efficiency and accelerate convergence. Simulation results show that the
proposed protocol outperforms existing schemes. It not only achieves near-optimal throughput with
faster convergence in static networks but also maintains strong robustness and adaptability in dynamic
environments. |
|
View Full Text
View/Add Comment Download reader |
| Close |
|
|
|
