Projects - IoBT REIGN CRA

Research Area 1: Agile Synthesis (Composition)


Future missions will exploit IoBTs made of thousands or tens of thousands of blue/military, red/adversary, and gray/citizen nodes, with a wide range of capabilities (Figure 7): from tiny occupancy sensors to drones with threedimensional Radar and LiDar sensors; from small on-board compute devices to powerful edge clouds with GPUs; and from actuators capable of modifying the environment in some way, to humans with powerful (albeit biased) perception, cognition, and action capabilities. Furthermore, future missions will need to be exceedingly agile: mission goals and needs may not be known until just before mission execution, and mission planners may not be able to (without the aid of automated tools) recruit and construct, at short timescales, IoBTs with sufficient resources to satisfy mission needs. The large scale of IoBTs will imply continuous churn, so discovery and composition solutions will need to be robust to failure or removal of assets as a normal operating regime.

In this research area, we aim to develop methods and fundamental limits underlying the recruitment and composition of IoBT resources, including potentially adversarial ones, into composite assets with sufficient sensing, compute, and communication capacities to satisfy mission needs and constraints, while at the same time multiplexing individual assets to achieve high return on investment. Recruitment, composition and reconfiguration of such assets must meet two fundamental needs. First, it should be possible to assemble (or re-assemble, for example, upon damage) composite assets comprising an IoBT of possibly 1,000s to 10,000s of nodes on demand and within minutes, if needed, despite high component heterogeneity, large scale, and presence of adversaries. Second, the aggregate properties of the composite, including timeliness, performance/functionality, security, and dependability, must be formally assured in an appropriately quantifiable and operationally relevant manner, subject to well-understood assumptions..