![]() |
SEVENTH
ANNUAL |
|
Multipath Signal Processing for Over-the-Horizon
Skywave Radar Target Localization Jeffrey Krolik and Richard Anderson Department of Electrical and Computer Engineering Duke University P.O. Box 90291/184 Engineering Bldg. Durham, NC 27708-0291 tel: (919) 660-5274 email: jk@ee.duke.edu Michael Papazoglou Naval Research Laboratory Abstract This paper concerns methods for
improving the performance of high frequency skywave over-the-horizon radar by exploiting
realistically attainable knowledge of complex multipath propagation conditions. Two
different problems are addressed here: (1) matched-field estimation of aircraft altitude,
and (2) coordinate registration and mode linking for multi-target latitude and longitude
estimation. Multi-dwell matched-field altitude estimation (MFAE) works by modeling the
changes in the coherent, but unresolved, direct and surface reflected target returns from
dwell-to-dwell. In particular, the pattern of rapid shape changes due to aircraft motion,
seen as fading of the complex radar return in delay-Doppler space, are strongly dependent
on aircraft altitude. Using a raytrace multipath propagation model to predict rapid
fading, and first-order Markov modeling to handle slow fluctuations due to target aspect
changes and medium fluctuations, a maximum likelihood estimate (MLE) of aircraft altitude
is developed. Real and simulated radar data indicates an altitude accuracy of better than
2500 ft at ranges out to 2300 km can be achieved. The second problem addressed here
concerns associating resolved multipath arrivals from different layers of the ionosphere
with the correct target and raymode type. This is critical to converting multipath time
delays measured by the radar with the target's ground range. The process of associating
slant tracks with raymode types and targets is commonly referred to as "mode
linking." Mode linker performance depends on both the accuracy of the ionospheric
model as well as the radar's slant tracker. The approach taken here is to use a
statistical model for ionospheric uncertainty in the mode linking/coordinate registration
algorithm in order to improve its robustness to environmental mismatch and hence minimize
the probability of incorrect mode linking decisions. Results for real and simulated radar
data are presented which indicate that the new technique can achieve better than a 2:1
improvement in target localization over conventional coordinate registration/mode linking
methods. |
![]()
Direct comments and questions to: webmaster@ll.mit.edu
© MIT Lincoln Laboratory. All rights reserved.