Bayesian Inference Techniques to support System-Level Post-Intercept Assessment (BPIA)

In this Phase I SBIR, Daniel H. Wagner Associates will construct a high-level, physics-based, probabilistic model of an anti-ballistic missile intercept event, characterizing and integrating various sensor phenomenologies from a probabilistic perspective; and we will develop multi-sensor data fusion techniques to estimate the effectiveness of the engagement. Our approach will emphasize Bayesian inference techniques to support System-level Post-Intercept Assessment (BPIA) by developing a Bayesian Network (BN) whose structure encodes the relevant ground truth features/signatures and their probabilistic relationships. Our sophisticated multi-sensor data fusion engine software, which employs our Bayesian Inference Engine (BIE) for feature-aided data association, will be configured to use this BN to process multiple diverse sensor outputs during and after an interception event. Further, we will develop supporting processes for translating high-level sensor data as evidence for the BN, to be processed alongside kinematic clues from the association/tracking components of data fusion system. Finally, we will identify and create a plan for implementing the more sophisticated BIE enhancements that would be undertaken during Phase II as part of a more complex BPIA prototype. Approved for Public Release | 16-MDA-8917 (15 November 16)