An Intelligent Routing Protocol for TDMA Mobile Ad hoc Networks Based on Deep Learning
Qiang Li, Jingchuan Sun, Yixin Li, Xunwei Zhao, Chunling Zhang, Zhigang Wang, Jie Bai, Jianguo Zhang · 2024
Mobile Ad Hoc Networks (MANETs) possess several unique qualities, including multiple nodes, dynamic topology changes, decentralized operation, and adaptability. However, conventional routing protocols employed in MANETs often suffer from various drawbacks, such as imbalanced network load, limited network throughput, inadequate resilience, and subpar transmission reliability. This article designs a TDMA mobile ad hoc network routing decision protocol based on deep neural networks (DNNs), which can provide different routing strategies under the superposition of multiple business type requirements and complex link conditions. And the routing process includes neighbor relationship establishment, topology flooding, intelligent routing strategy decision-making, and routing establishment. After learning topology information and service demand, nodes can calculate the routing table and find the most suitable path. Finally, the routing protocol is developed on the OMNeT++ simulation platform, and its effectiveness and excellent network performance are tested through experiments.