Optimization of Sensor Management/Sensor Registration

To provide multiple layers of defense against increasingly complex threats, the heterogeneous sensors of the Ballistic Missile Defense System (BMDS) must be carefully coordinated to manage their load and make optimal use of their diverse capabilities. The approach to sensor resource management (SRM) must address a number of key needs including: sensor dynamics and their limitations, uncertainties in sensor and target behaviors, responsiveness to a changing scenario, and the ability to prioritize according to the importance of engagement needs such as discrimination and fire control quality tracks. To address these issues, Numerica proposes the development of a partially observable Markov decision processes (POMDP) framework for sensor resource management. Such a framework provides a flexible modeling capability, is non-myopic, addresses uncertainty in a changing dynamic environment, accommodates runtime controls, and, due to recent advances in algorithm development, is computationally feasible. Numerica will prototype and test its SRM system in the BMD Benchmark and will work with Raytheon to improve the design as it may work with future sensors. The algorithms will be developed initially in MATLAB, followed by a design and implementation in JAVA or C++.