Information fusion algorithms and analysis for an exemplar detection of intent problem

Abstract

Information fusion algorithms for data association and inference are applied to a representative intelligence gathering problem in which signals of intent are monitored by multiple imperfect sensors over a period of time. Two sets of algorithms are developed: a brute force set which makes best use of the data but is not efficient, and an approximate set which sacrifices some performance for efficiency. The algorithms are applied to simulated data to generate evidence of intent. Then a Monte Carlo process and an associated metric are developed to evaluate the performance of the algorithms under different levels of uncertainty in the data. This analysis helped to validate the algorithms and it can also provide useful system design guidelines.