Statement of Research Interests
My principal research interests are in communications, signal processing and applied harmonic analysis. Most of my research has been directed toward solving engineering problems for government partners (MITRE), private industry (Radyn, Inc.), and university-linked government contracts (Opportunistic Sensing ARO-MURI).
My work with MITRE has spanned several decades and is principally concerned with underwater acoustics and telecommunications. I created several detection and tracking algorithms to solve specific acoustic analysis problems as well as implementing advanced harmonic analysis techniques to analyze unique acoustic signatures. I produced several papers on vibration analysis, time-frequency transforms, wavelet analysis, and neural networks. More recently, I have been working on systems to improve Radio Frequency (RF) communications with a network of acoustic sensors. This involves taking cellular and satellite communications technologies and applying them to complex and congested communications settings where their behavior has not been previously studied. I am currently co-authoring a paper where spread spectrum techniques are analyzed for a field of naval sonobuoys communicating with an aircraft. MITRE, in conjunction with the Office of Naval Research and the Navy, will be conducting field test of our algorithms.
My work with private industry has focused on communications engineering. For Radyn, I have invented and implemented several algorithms and heuristic methods for solving difficult spectrum allocation problems. These include channel planning and prioritization algorithms within Radyn’s Poseidon™ software and frequency search and propagation algorithms within their Openlink™ tool. I continue to update and maintain the software on an as-needed basis to help deal with changes to the FCC standard procedures and the changing requirements of the spectrum allocation environment. I have also conducted research and produced several smaller papers presented to the FCC and industry on various spectrum management topics including some on spectrum sharing for mobile satellite systems and interference analysis for terrestrial mobile bands. One of these was presented to the FCC in the summer of 2012.
My other industry work has included 4 MIPS grants, two with Prof. Tretter and Telecontinuity Inc. to create a disaster-proof telecommunications network and the other two with the Norbert Wiener Center and ReoSYM Inc. to investigate data compression for innovative cellular basestation design.
This latter work led to several interesting results extending existing algorithms for dimension reduction. John Benedetto and I saw connections between this and work being done on the Opportunistic Sensing ARO-MURI that he, I, and others (including Prof. Chellappa and his team) have been working on. We are creating a mathematical theory for interpreting opportunistic sensing results using the theory of finite frames. We term this technique Reactive Sensing. The idea is to interpret the partially redundant output of multiple sensors as a frame for the resulting sensed data space as opposed to the usual notion of only considering the output of the “best” sensors which results in a basis representation. We have been working for the last nine months to build the mathematical framework for this theory. Among other results, we have shown that given certain assumptions about overlapping sensor characteristics, it is possible to reconstruct sensed data even when one (or possibly more than one) sensor fails. The result hinges on constructing a frame for the sensed space based on the entire set of sensor outputs.
In the future, I plan to focus on continuing the development of Reactive Sensing. In particular, I believe useful results will emerge when we consider what happens if the SNR of the sensed signals varies for different sensors. I believe the results will provide concrete computations that will realize the trade-off between the SNR of the redundant sensed data and the degradation of the probabilities of detection and false alarm for the sensor network.
I will also continue working on interesting communications problems that arise at MITRE. The current demand for spectrum is likely to create many challenging problems as older less efficient technologies used by MITRE’s clients must be updated to work in more congested interference environments. I hope to continue to make contributions to this effort.