Data association and geolocation for electronic support systems

Abstract

Data association (DA) in highly sensitive electronic support (ES) systems is nontrivial when closely-spaced, low probability-of-intercept (LPI) emitters, in cluttered radio-frequency (RF) environments, are sought. A practical processing framework for these situations is presented in this paper. Joint probabilistic data association (JPDA) is used to handle the multitarget tracking (MTT) problem in frequency-augmented direction-of-arrival versus time coordinates (i.e. ‘signal’ space). It is shown that the JPDA update has a pleasantly simple form when processing a serial stream of time-stamped measurements. A Bayesian M-out-of-N track confidence model is derived using the beta distribution and used to ‘integrate’ automatic track management functions (i.e. JIPDA). A maximum-likelihood expectation–maximisation (ML-EM) algorithm is also derived and applied to perform bearing-only target-motion-analysis (TMA) in geographic coordinates for confirmed JIPDA tracks. The application of EM compensates for previous measurement-to-signal-track assignment errors and decreases the deleterious influence of spurious measurements, due to clutter or other emitters, without combinatorial complexity. Highlights: The ES data processing problem is factored into signal and geographic coordinates. JIPDA is applied to solve the multi-target tracking problem in signal coordinates. Beta distributions are used to yield a Bayesian M-out-of-N track confidence model. ML-EM eliminates outliers and locates emitters in geographic coordinates.