Intelligent Cloud-Based Hybrid Cellular and Mobile Ad Hoc Networks for Future Live Training

Future Army live training systems will be based on cloud computing with portable user equipment. In this project, we propose an intelligent three-tier cloud-based hybrid cellular and mobile ad hoc network architecture with a set of advanced networking and edge computing techniques, which enables high-bandwidth low-latency data delivery and real-time data processing to support future live training simulation capabilities. Specifically, i) a scalable hybrid cellular and ad hoc device-to-device wireless network with efficient proxy discovery and routing protocols is designed to significantly enhance the throughput of data collection and distribution for training participants; ii) edge servers with machine learning based dynamic task offloading algorithms provide low-latency data computing and training simulations at training sites; iii) intelligent resource management and data delivery schemes with distributed cloud data centers and edge servers allow fast multi-source multi-path training data retrieval, caching, and distribution; iv) deep reinforcement learning based transmission scheduling and control schemes optimize network delay performance and energy efficiency under dynamic environments. An integrated network simulation platform will be developed to conduct evaluations. The proposed schemes can be implemented in software without changing commercial network infrastructure. They will significantly improve network bandwidth, reduce latency, and enhance data delivery efficiency for future training.