Advancement: RF Emitter Geolocation using a Single LEO Satellite

Speaker Name: 
Patrick Ellis
Speaker Title: 
PhD Student (Advisor: Farid Dowla)
Speaker Organization: 
Electrical Engineering
Start Time: 
Wednesday, November 28, 2018 - 12:00pm
End Time: 
Wednesday, November 28, 2018 - 2:00pm
Engineering 2, Room 553
Farid Dowla

Abstract:  This thesis addresses the important problem of the high-accuracy geolocation of a ground-level RF emitter based on a weak signal received on a fast-moving single LEO satellite receiver platform. The proposed technology can be expected to have a significant impact on search and rescue operations, spectrum monitoring, and environmental tracking. By analyzing the technical challenges of real-time single-pass solutions over a wide area search space, in this thesis it is shown that a high-accuracy target location is feasible knowing only the center frequency and modulation of the RF emitter.

The LEO platform employs a novel geolocation algorithm that combines the spatial diversity provided from the platform’s high velocity with information about the earth’s rotation and surface that maps the received signal to an emitter location. The geolocation algorithm exercises optimal error minimization algorithms based on Unscented Kalman filters (UKFs) that is resilient to initialization, ephemeris and oscillation errors. Performance bounds in the form of the recursive Posterior Cramer Rao Bound (rPCRB) have been computed and adapted to the single sensor scenario – giving a measure of the lowest mean squared error that a nonlinear iterative estimation algorithm can achieve.

A complex and realistic simulation environment has been developed and the code produces data for analysis and evaluation. We show close agreement between derived theoretic performance bounds and simulated results. Geolocation optimality in terms of signal center frequency, coordinate systems, and flight paths is studied and analyzed. Raw IQ data from the CYGNSS satellite constellation has been obtained for final validation.