Sensor Data Fusion for Target Classification and Identification

Toyon Research proposes to develop a fusion module that fuses measurements collected by any number of sensors over time. The fusion module will consist of a Bayesian Network and a statistical database. The Bayesian Network will update beliefs about whattype of target is being measured by taking into consideration a priori information regarding the types of airborne objects in an area, the kinematic state estimate provided by a tracker, and the capabilities of the sensors providing measurements. Thestatistical database will provide information regarding the expected characteristics of the sensors and the expected kinematic characteristics of each target type. One of the important advantages of our approach is that information from sensors whichprovide data at different levels in a target-class hierarchy can be effectively fused. We will build a computer simulation of the non-cooperative target recognition (NCTR) architecture so that we can test the fusion module using a realistic number ofsensor measurements. Optional tasks include the analysis of ATR outcomes for building the statistical database, installing and testing the fusion module within a government facility, and integrating the fusion module with a tracker and sensor tasker.