Sensor Allocation in Highly Mobile Networks

"Under the proposed project, Advanced Ceramics Research (ACR) and the University of California Los Angeles (UCLA) will team to evaluate the utility of learning methods as applied to sensor management. Reinforcement learning methods suitable to control thesensor network for the discovery of remote Threat Locations and Radio Jamming frequencies (TLRJ) will be developed. In conjunction with RL methods, the team will develop optimization techniques that compute the best Vector of Sensor Positions (VSP) andVector of Sensor Radio Link Frequencies (VSRF) given threat locations and detected radio frequency jamming signals. Sensor management algorithms developed at UCLA will be integrated with the Smart War-fighter Array of Reconfigurable Modules (SWARM)developed by ACR. In Phase I the team will demonstrate the management system with actual flight testing and evaluate reinforcement learning for the discovery of TLRJ information. A successful Phase I will lead to a Phase II program where sensors areautonomously controlled under the direction of VSP and VSRF solutions. This system will be applicable to existing UAV platforms and will provide significant benefit to next generation high-speed UAV platforms with aggressive maneuver capabilities.Commercial markets include application in autonomous control of Navy SWARM UAV's and air traffic control for collision avoidance as well as congestion management."