Prof. Fowler's Research                                                                

Emitter Location Research Group

Director: Prof. Mark L. Fowler

Participating Faculty: Prof. Eva Wu, Prof. Edward Li

Participating Post-Docs: Mo Chen

Binghamton University’s Emitter Location Research Group (ELRG) focuses on the application of signal processing theory to improve the ability to passively locate transmitters.  The need to locate emitters arises in both military and commercial settings; the latter being in response to the FCC’s mandate that cellular providers have the ability to locate cellular phones used to place 911 emergency calls.


Sources of Research Support:


Main Areas of Research:


1. Multiple-Platform Methods – where signal data is intercepted at several widely separated platforms and shared between the platforms to locate the emitter.  We focus mainly on the use of time-difference-of-arrival (TDOA) and frequency-difference-of-arrival (FDOA).


Data Compression (Fowler):

A key impediment to the TDOA/FDOA method is the time it takes to transfer the data from one platform to the other.  We are developing ways to reduce the amount of data that needs to be sent.  By exploiting characteristics of the signal to be exploited together with the characteristics of the TDOA/FDOA processing, we have shown that it is possible to significantly increase the compression ratio.  We are developing ways to exploit these characteristics by using the wavelet transform.

We have successfully defined new distortion measures and have shown them to improve the compression ratio while negligibly impacting the TDOA accuracy.  We are currently exploring the use of multi-objective optimization to allow us to trade-off between the requirements for TDOA accuracy and FDOA accuracy.

For the specific case of locating radar emitters, we have developed a method that gates the pulses into a matrix and then operates on this matrix using the SVD to achieve compression ratios from 20:1 to over 100:1 (depending on the signal characteristics).


Multipath Mitigation (Fowler and Li):

            One of the largest sources of error in a TDOA/FDOA system comes from the impact of signals arriving on paths other than the direct path from emitter to receiver.  These multipath errors perturb the peaks on the cross-correlation  that are used to estimate the TDOA/FDOA values.  By exploiting the fact that you have access to data that has traveled over multiple multipath channels, it is known that the channels can be  blindly identified without knowledge of input signal statistics by doing a least-squared based approach.  What has not been assessed is the suitability of such blind equalization for TDOA/FDOA estimation.  Furthermore, once the task of  equalizing the channel for emitter location is significantly different from the case for communication because for emitter location the things of interest (time delays) are part of the channel!  We are developing so-called partial equalization methods for this case.


2.  Single-Platform Methods – where signal data is intercepted at a single platform over a time span of several tens of seconds and used to locate the emitter.  We focus on the use of frequency measurements and long-baseline-interferometers, although other methods are considered as well.


Application of MEMS Accelerometers (Wu and Fowler ):

            Existing methods are sensitive to errors in the knowledge of the time history of the receiving antennas’ position and velocity.  We are exploring ways to use microelectromechanical systems (MEMS) accelerometers placed at the antennas to reduce the uncertainty in the antennas’ position and velocity and hence improve the achievable location accuracy.  We have derived algorithm structures exploiting federated filtering to allow the MEMS-derived data to be optimally combined with the on-board navigation data without requiring any redesign of the navigation system.  We have derived extremely tight bounds on the accuracy attainable using such techniques and have shown that a 100 mg MEMS accelerometer would provide at least a 60 dB reduction of local velocity estimation error assuming a 2m/s 1-s local velocity uncertainty in the absence of the accelerometer.


Algorithms Less Sensitive to Antenna Uncertainty (Fowler):

            As an adjunct to reducing the uncertainty, we are working to develop location algorithms that are less sensitive to the antenna uncertainties.  These methods strive to exploit variations of the total least-squares method.


Use of Digital Terrain Data (Fowler)

            It is hard to accurately estimate the emitter's altitude using single platform methods.  We have been exploring ways to exploit digital terrain maps to improve the altitude estimate - a side benefit of this approach is that it also gives significant improvement in the X-Y estimates, too.  We have developed theoretical results showing the impact of digital terrain data and have developed and tested algorithms that exploit such terrain data.




Matlab Files

Back to Prof. Fowler's Homepage