Doppler geolocation with drifting carrier

This paper describes a particle filter approach to improve passive Doppler geolocation of non-cooperative radio frequency (RF) emitters from measurements collected on-board a single moving sensor. Existing Doppler-based geolocation techniques assume that the emitter´s carrier frequency is unknown but stable over the course of the triangulation. However, this assumption is often violated by today´s commercial devices that have manufacturing tolerances that allow significant carrier drift, and the application of static optimization procedures to this non-stationary state description often leads to nonsensical answers. This paper proposes a recursive approach based on particle filters that tracks the non-stationary carrier drift while discretely representing the emitter´s multi-modal state space through time to ultimately converge to the true emitter location. The new approach is compared to Newton optimization and extended Kalman filter (EKF) based solutions.