CMU

Research

My current research interests include:

  • Dynamic Spectrum Access: Developing new models of spectrum management and methods for enabling coexistence among radio services. Of particular interest is coexistence with legacy systems, including interoperability and compatibility between federal and non-federal systems.I am developing new technical models and policies to address legacy systems and emerging new services. I also plan to explore methods of reallocation based on coexistence, cooperation and/or tolerance.

  • Wireless Security: Developing models to address security concerns in emerging dynamic spectrum access systems. Defining risk attributed to new dynamic spectrum access technology. Exploring jamming and anti-jamming mechanisms through dynamic spectrum access technology.

  • Network/Telecommunications Policy: Implementing data-driven decision-making; Adopting process transparency and accountability; and Reforming radio spectrum management.

Below is a set of example topics I would like to explore:

  • Is it possible to convert customary transmitter-power constraints into receiver interference limits? This is the question of, "how do we move interference issues from the transmitter to the receiver?" What method might we consider for negotiating towards maximizing social surplus?

  • Would a spectrum model based on interference limits offer any advantage in terms of improving spectrum use -mitigating conflicts, maximizing utility?

  • What interference threshold models are acceptable? Traditionally conservative thresholds are set at a high false positive; interference is strongly dependent on the starting assumptions (metrics, measurement, thresholds, models...). How might other thresholds be justified?

  • Does beamforming actually reduce a threat in the face of rogue jamming devices? What are the limits of high efficiency in spectrum reuse from beamforming in live radio environments?

  • Develop an initial test-bed facility to study and investigate the possibilities of shared spectrum access between disparate users. The desire is to build upon this capability and make it a repeatable process so that collected information, lessons-learned, and its applicable infrastructure can be re-used for future experimentations and/or demonstrations. It is envisioned to expand this capability into a transparent, comprehensive and distributed testing & demonstration environment that seeks to leverage current and future technologies as a means to model, simulate and/or perform live experimentation. Can a red team / black team approach be used to assess spectrum testbeds? Establish two teams that are intended to focus on different sides of a spectrum problem.

  • What enforcement and monitoring programs make sense for the FCC in the future? There are at least two very different areas to consider – future of broadcast decency issues AND spectrum enforcement – and the models for each will likely diverge.

  • Is there a future model of dense small cells in urban areas that may be community/end-user deployed, which is not solely controlled by, and may be independent of, carriers? What would it take to enable? Why can't I have a home router with Sprint, VZW, AT&T, WiFi, 3.5GBps, etc. radios that I (as an end-user) can manage? What complementary infrastructure (poles with power for antennas and backhaul siting, maybe collocated at anchor institutions) is needed to have to make this work? How can this alternative model of community-level wide-area (i.e., not just in my home) wireless connectivity grow and interconnect with provider models?

  • What is a functional mixed model of spectrum regulation? One could look at legacy spectrum allocation as a very static database approach to spectrum management (the allocation tables are the dbase and it changes slowly and at great expense), versus a full-blown cognitive radio vision as sensor-enabled handsets that negotiate spectrum access in real-time. Reality is somewhere in between. What are the economics of this future and what has to change in technology, business and policy to allow for it?

  • How much to sense and sample in cognitive radios? Examine the problems and performance of wideband spectrum sensing methods (Nyquist and sub-Nyquist sampling sensing methods) under situations of high Doppler and time-varying characteristics of links caused by high relative mobility between nodes, as well as the multipath effect caused by complex terrain environment.

  • What is the role of compressive sensing in cognitive radios? In wideband multi-channel networks, could one use compressive sensing-based channel estimation and signal detection methods to alleviate processing overhead and the requirement of ADC, DSP or FPGA, to achieve the Space-time-frequency spectrum reusing and dynamic spectrum management. Because the sensing object is the signal with a known modulation, we might not need to accurately reconstruct the signal, rather just extract the characteristic parameter of wideband wireless signal, and then determine the channel and what modulation mode is used with a much smaller processing overhead.

  • How might we improve the adjacent channel selectivity for whitespace receivers in a cost effective manner? How selective do these devices really need to be given the current and future legacy receivers?

  • A few thoughts on research relating to broadband access: Is it possible to create inexpensive, simple and mutable broadband access edge equipment? How might Software Defined Networking be used as a tool to create and manage simple broadband access equipment (wireless AP, cable modems, optical terminals…) based on technology that can allow the network edge to evolve and adapt. Can we use NetfFow and/or other data sources to assess interconnection conflicts? Examine the role of empirical data analysis of traffic flows across the Internet. Can we use BGP as an IGP for cloud and data center routing? Examine the application of policy based routing in environments with diverse application and content sources. I’m interested in the many issues surrounding privacy on networked systems. I’m interested in understanding how to better inform policy makers with the right set of data and analysis of that data, “data driven policy”.

  • I also work on applying decision analysis and cost benefit analysis to problems in the ICT policy space. These problems, such as network interconnection and the provision of societally beneficial (accessibility or emergency) services, are complex in terms of incentives and associated cost or pricing issues.

  • An arguably less academically rigorous interest of mine involves helping policymakers understand the technical and economic aspects of networked systems. Sometimes this work can take the form of technical tutorials (e.g., how a technology works) and other times it can involve diving deeply into the potential problems that might arise because of policy decisions that don’t appreciate the implications of an emerging technology.