Program Manager: Mark
A. Neifeld, Ph.D.
The Sensor Topology for Minimal Planning (SToMP) program leverages high-dimensional mathematical insights to create new DoD capabilities. These capabilities capitalize on emerging opportunities where sensors are miniaturized, pervasive, and coordinated. By developing new mathematical methods and techniques, the program seeks to solve a multiplicity of network and sensing problems that arise in a variety of military situations. Much current technology forces real world problems, which intrinsically possess many degrees of freedom, through a one-dimensional network where the mathematical formulation has, by necessity, lost crucial information. By injecting decades of topological and geometric advances into the analysis, the program will revolutionize how networked sensors and autonomous agents are analyzed, distributed, and controlled. Of central importance will be a systematic determination and exploitation of minimization of total sensor complexity. More precisely, given a mission having a variety of costs to be optimized (e.g., total number of sensors, network bandwidth, or power consumption) the program seeks to determine solutions that require the least resources with respect to such metrics. Thus, the program aims to derive optimal solutions in that it focuses on implementations that are minimal in terms of reducing the total sensor complexity to the minimal configuration required.
The fundamental mathematical and computational tools developed and implemented in this program will have broad impact across several avenues of defense applications as the sensor networks, autonomous systems, and configurable sensor platforms amenable to the proposed analysis reside throughout the military. Examples include fast distributed algorithms for sensor coverage, target encirclement, identification, and tracking in coordinate-free networks; pursuit and evasion in non-convex domains with many degrees of freedom, including visibility-based pursuit in non-convex three-dimensional regions; and landmark-based navigation algorithms for mapping and localization in systems with minimal sensing capabilities, as well as incorporation of stochastic features into all of the above.
|
