Hybrid Model Based/Knowledge Discovery Approach to Situation Assessment

This proposal explores a hybrid model based/knowledge discovery approach to situation assessment (SA), especially suitable for detecting and identifying targets in urban network centric warfare (NCW) environments. The approach recognizes significant enemy activity patterns from messages collected at distributed NCW nodes (e.g. sensors and various units) by taking into account environmental clutter. It uses clustering to perform a space and time-series analysis of messages without requiring semantic information. The approach, for example, will be able to detect spatially correlated moving units over time within the environment. Detected patterns trigger the need for assessing newly developed situations, resulting in invocations of various doctrine-based computational models, including causal static and dynamic Bayesian belief networks (BNs) considered in this proposal. The selected models then perform SA based on other observables propagated as evidence into the models. Our approach extends further by recognizing significant patterns without relying on doctrinal knowledge. It is based on Latent Semantic Indexing (LSI), which is a proven technique in text based information retrieval applications. We use LSI to extract underlying patterns from observables reported in formatted or text messages. These patterns form a "normal" profile against which incoming observations are matched so as to detect any unusual activities.