Leveraging Digital Twins to Enhance Performance of IoT in Disadvantaged Networks

The U.S. Army is investing in research to deliver enhanced sensing, situational awareness, computing, and decision-making capabilities via IoT devices on the battlefield. One of the current shortfalls with existing capabilities involves the use of large wireless mesh networks to support disadvantaged users. Networks, like the Soldier Radio Waveform (SRW) suffer from increased latency and packet overhead that affect system performance and severely limit the operational employment of the system. In this paper, a novel hybrid topology management technique using a digital twin of the propagation environment and rapid assessment, cluster creation, and node assignment process has been proposed and investigated to show Software-Defined Networking and K-Means Clustering machine learning have potentials for increasing the number of users able to connect to wireless mesh networks. Preliminary results show that this technique can reduce the path length in large graphs at the expense of edge connectivity while delivering a level of service to all users that was previously unavailable.