Previous Abstract Return to Session D3 Next Abstract

Session D3: PNT Situational Awareness: Sensing Technologies 1

PNT-SENTINEL: Enhancements to Proliferated LEO PNT-SA
Scott Jones, Aaron Pung, Slingshot Aerospace
Location: Ballroom C
Date/Time: Monday, Jun. 2, 2:10 p.m.

The Position Navigation and Timing (PNT) Secure Electronic Navigation Threat Intelligence and Location (PNT-SENTINEL) is a Direct to Phase II SBIR effort to enhance the capabilities of the Slingshot Data Exploitation and Enhanced Processing (DEEP-PNT)[1] prototype. DEEP-PNT ingests, processes, and exploits GNSS observations recorded by receivers onboard proliferated Low-Earth Orbit (pLEO) commercial satellites to generate near real time globally persistent PNT Situational Awareness (PNT-SA) data. The enhancements developed under PNT-SENTINEL include moving beyond GPS to multi-GNSS observations, improved RF models for detection and estimation, and prototyping spatiotemporal change detection methods.
Currently, the DEEP-PNT user interface generates and displays global heatmaps of probable interference in the GPS L1 band as inferred from observables from pLEO GNSS receivers. PNT-SENTINEL aims to enhance the existing effort in three ways. Previous experiments using manually tasked on-orbit sensors showed that in regions in which GPS L1 interference is consistently observed by the DEEP-PNT system, interference is present in other bands as well. Improvements to the receivers onboard the latest generation of commercial pLEO satellites has enabled multi-GNSS observables collection. PNT-SENTINEL extends the GPS L1 PNT-SA capabilities developed under DEEP-PNT to the GPS L5 and Galileo E1 and E5A civil signals.
PNT-SENTINEL integrated more sophisticated signal propagation and receiver models, and detection algorithms, to improve interference detection accuracy over l baselines demonstrated in DEEP-PNT. Path loss correction is accomplished by ingesting published GPS ephemeris, while improved receiver antenna models and attitude information enable correction for angle of arrival. Although individual observations are coarse, the large scale of the dataset means that spatially aggregating the data along with improved predictive modeling and Neyman-Pearson statistics developed for the residual values greatly improve interference detection accuracy. In addition, this enables estimation of metrics including the received interference power, and consequently, the jamming to signal (J/S) ratio. Initial experiments have shown that regional biases in the expected received power residuals output by the enhanced models are consistent with results on self-interference from intercompatible signals reported in the recent literature [2].
Feedback from operators using the DEEP-PNT tool stressed the need to develop automated change detection of the global signal interference environment. Preliminary investigations of the observation data have shown temporal periodic structures in regions with known persistent interference, as well as changes in the interference environment corresponding to known world events. Initial experiments have demonstrated the feasibility of change detection using simple statistical methods. Using these results as a baseline, the PNT-SENTINEL team is actively exploring AI/ML methods. Prototype AI/ML change detection algorithms will be validated against known GPS test events in CONUS published as notices to airmen (NOTAMs).
[1] Jones, Scott., Shagnea, Peter., Fiske, David., Pung, Aaron., “DEEP: A Proliferated LEO PNT Situational Awareness Threat Monitor,” Proceedings of the 2023 Joint Navigation Conference (ION JNC 2023), San Diego, California, June 2023.
[2] Hegarty, Christopher J., O’Hanlon, Brady W., "Enhancement and Extensions to a Cyclostationary GNSS Self-Interference Model," Proceedings of the 37th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2024), Baltimore, Maryland, September 2024, pp. 804-817. https://doi.org/10.33012/2024.19734

Previous Abstract Return to Session D3 Next Abstract