Quick Look

Using an HF Time Difference of Arrival Geolocation Network To Perform
Direction Finding, 16-R9563

Printer Friendly Version

Principal Investigators
Brent Fessler
Don R. Van Rheeden

Inclusive Dates:  08/01/05 – 12/01/05

Background - Ionospheric propagation of HF signals presents a challenging environment in which to perform radio direction finding (DF). Multi-hop propagation, random time-varying polarization, and other effects tend to degrade HF skywave signal DF estimator performance. Multiple ionospheric propagation modes also make accurate HF skywave signal time-difference-of-arrival (TDOA) geolocation difficult, if not impossible, to perform. For receivers separated by hundreds of kilometers, an HF signal can diffract off of different layers in the ionosphere and travel to the receivers along widely varying propagation paths. However, if the TDOA receivers are in the same geographic area, then there may be conditions in which an HF skywave signal will propagate along a common path or a common set of paths to each receiver. If such conditions can be identified, it may be possible to use close proximity TDOA systems to perform HF direction finding.

Approach - A DF from TDOA approach used by this project was a generalization of a time- delay DF approach developed by Berdugo, et al. [B. Berdugo, M. A. Doron, J. Rosenhouse, and H. Azhari, "On direction finding of an emitting source from time delays," The Journal of the Acoustical Society of America, Volume 105, Issue 6, June 1999, pp. 3355 - 3363] for acoustic signals. To test the TDOA DF concept, four HF TDOA systems were deployed around San Antonio and used to collect signal samples on several HF signals. These systems were separated by distances of one, five, seven, and ten kilometers.

Accomplishments - From September through November 2005, several signal collection exercises were performed. These collections were performed over a frequency range from 8 to 15 MHz. Transmissions collected included AM broadcast stations, amateur radio operators, and known check targets such as WWV and CHU. DF results on these transmitters were mixed. The figure below shows one case where the approach using three of the four TDOA sites produced an accurate DF estimate – note the nearly parallel hyperbolic asymptotes. The transmitter was an AM broadcast at 14.192 MHz. In other cases, the hyperbolic asymptotes were not parallel, and the DF estimates were inaccurate. This inaccuracy could be a result of HF skywave signals having random time-varying polarization, TDOA estimation errors, and the receiver baselines used for this study may be too large to ensure spatially diverse reception of a signal perturbed by common propagation effects.

In this example of DF results using three TDOA sites, the arrows represent the DF estimate for an AM broadcast signal.

2006 Program Home>