Efficient Sensor Management for Optimal Multi-Task Performance (ES-MaTe)

Knowledge Based Systems, Inc. (KBSI) proposes to investigate innovative methodologies and algorithms for Efficient Sensor Management for Optimal Multi-Task Performance (ES-MaTe), The goal of the proposed research is to investigate an efficient sensor resource management method for sensor scheduling specific to tasks required by a missile defense system. The Ballistic Missile Defense System uses multiple disparate and spatially separated sensors for multiple simultaneous tasks such as target search, detection, acquisition, discrimination, target tracking. Each of these tasks consists of minimization of uncertainty in estimation of some underlying stochastic process. However, sharing of resources with other tasks results in some degradation of the estimation performance. ES-MaTe will develop a constraint optimization technology for non-myopic multi-sensor scheduling that will ensure some level of minimum performance for each task by imposing performance constraints based on uncertainty measurements, while maximizing the total information gain. A feature of the optimization problem includes the imposition of several types of task specific sensor utilization constraints. The optimal non-myopic sensor scheduling requires dynamic programming that becomes unpractical due to complexity arising from combinatorial explosion. Approximate dynamic programming techniques will be investigated for practical implementation.