Preemptive Actions with Dynamic Anticipatory Targeting

We propose a system for real-time target identification and intent prediction of time critical targets (TCTs) from limited tracking data. Given the urgency of destroying or disabling an emerging target (e.g. a missile with a biological warhead), we focus on enabling preemptive action through predictive situation assessment. We assume a high degree of uncertainty about the quality and availability of track information needed for target identification. As new tracks emerge, the system initiates track identification via (a) navigation through a Decision Tree, accelerated by Branch Prediction algorithms, in search of valid and useful identity attributes, (b) preemptive, look-ahead invocation of networked Collection Agents to acquire additional critical information, and (c) probabilistic assessment of target identity and intent via a bank of doctrine based Bayesian Networks (BNs). The BNs assess various track attributes to progressively improve overall knowledge of target identification and intent. As the situation evolves, a Central Control Unit adaptively adjusts critical threshold levels controlling information collection to optimize the computational load, and quickly produces actionable decisions. Our overall objective is to produce a quick and agile system, which strives to increase the certainty of identification and intent of a TCT, from an emerging track with limited information.