Intelligent RF-IR Data Fusion using Artificial Intelligence Techniques

The ultimate goal of this proposed effort is to better utilize disparate sensor data to perform better correlations and lethality assessments through the development of a sensor data fusion system using artificial intelligence (AI) techniques and the concept of extracting features from raw sensor data and reasoning about those features and their hypothetical association with hypothesized objects. Building on the Phase I success, we will develop low-level sensor signal processing code and high-level feature extraction software to determine physical and kinematic features of the sensed objects. We will begin with the specifications for the primary sensors and descriptions of the primary targets of interest as well as the operational situations that are expected. An important objective of this Phase II effort is to gather additional experiment data to determine operational accuracy and limitations, validate our approach, and supplement existing databases. Using the existing and collected data as a basis, we can perform knowledge elicitation with our UDRI sensor experts to determine what features can be reliably extracted from the sensor data and in what situations. Based on what features are extractable, we can develop the automated reasoning to take advantage of them for both correlation and classification.