Networks

From fighting wildfires to cooperative robotic swarms, many research topics require understanding how to establish, leverage, and analyze networks between numerous and diverse agents.

MTRI performs future-forward research in a wide variety of network science domains—from traditional RF communications analysis to group formation and assessment on social media platforms. A special focus of the lab is to develop new hardware and leverage machine learning techniques to deploy robust communications networks in harsh environments, such as underground caverns or active wildfires. This selection of projects demonstrates our experience in a wide variety of application areas.

  • Bayesian Adaptive Robotic Control System (BARCS) - DARPA-TTO, Subterranean Challenge Program: Underground environments pose a particularly challenging arena for deploying autonomous multiagent robotic systems, but these systems would be immensely valuable for disaster response and defense application. Our team developed novel controllers for simulated multi-agent teams of autonomous agents to explore underground environments using distributed graph search algorithms and MANET technologies.
  • Structure Situational Awareness for Swarms (SSAS) - DARPA-TTO, Offensive Swarm-Enabled Tactics (OFFSET): We developed a synthetic swarm-based demonstration of see-through-wall floorplan reconstruction technology with an automated graph-based mission planner.
  • Predictive Social Network Graphing - Office of Naval Research: This project advanced the theory of graph convolutional neural networks by developing and implementing a predictive model to detect changes in social network graphs derived from Reddit data.
  • Public Safety Communications Research (PSCR) division of NIST: Wildfire DLN is a combined hardware/software solution designed to establish robust communication networks for first responders in areas with little or no prior infrastructure. The project leverages capability currently in use by the first responders and supplements with minimal additional hardware, thereby creating a low-cost solution to allow for seamless data transfer between areas with connectivity (such as a base camp) and those first responders who are on the front lines in areas with no connectivity. This work was performed in collaboration with Indiana University and the University of Tennessee.
  • Distributed Radar: The AFRL Study  explored the requirements and benefits of using networks of Radars for surveillance.  Instead of simple handoff from one Radar to another, the Distributed Radar program explored the exploitation of the simultaneous angular diversity afforded by the network for moving object imaging.
  • Shared Spectrum Access for Radar and Communications (SSPARC): This DARPA program explored waveform design and processing for a network of cognitive radios to simultaneously perform Radar and communications functions.  We examined waveform design, network timing requirements, image reconstruction methods, and the construction of networking signal packets to allow for the coordination of the network.