jump to navigation

Wireless Communications Research May 22, 2008

Posted by flashbuzzer in Research.
Tags: ,
1 comment so far

Recently I’ve been thinking about the current state of research in wireless communications. In particular, I’ve been pondering the following question: if I were a first-year grad student in wireless communications starting my Ph.D. this fall, what would be a good research problem to investigate?

Back when I entered grad school, I had a very hazy idea of what problems to tackle. My initial interest was in multi-input multi-output (MIMO) wireless systems. After poking at a multiuser MIMO beamforming problem for a while, one of my advisors suggested that I look at MIMO relay channels. That set me on my current research path, and I’ve come to regard cooperative communications as a neat and interesting research topic.

After recently attending VTC2008-Spring, I want to throw out the following three general topics that should be interesting and relevant for the next 5 or so years.

1. Cognitive radios: Given the ongoing whitespaces debate, where the FCC is being pushed to allow greater freedom to unlicensed transmitters, I think that this will be a timely and relevant problem for several years. There are numerous open problems in this area, particularly in the information-theoretic realm (since a cognitive radio channel can be regarded as a special case of an interference channel, and some progress has only recently been made on characterizing the performance limits of the interference channel). From attending some cognitive radio talks at VTC, I’m now convinced that other interesting problems lie in this domain, such as how to perform spectrum allocation and how to properly price available spectrum. Perhaps game-theoretic tools can be applied here?

2. Vehicular communications: This area also contains a raft of pressing technical issues. A nice motivating example for studying vehicular communications is as follows: suppose that you are driving along a road and an accident occurs up ahead (but out of your view). Ideally a signal should be sent from the wrecked car to your car, warning it that an accident has occurred ahead and advising you to slow down and/or take a suggested alternate route to your destination. Numerous auto companies have active research programs in this area, including Daimler and Tata Motors. A nice repository of information on vehicular communications can be found at this EPFL link, and a brief perusal of that page indicates that security and privacy issues are also of paramount importance; for example, it would be highly undesirable for a hacker to send a rogue signal to a moving vehicle, causing it to turn and crash into a brick wall.

3. Applications of biology: I used to regard biology as a subject unworthy of my attention (probably due to the relative lack of mathematics involved). Then I attended two interesting talks at VTC where biological principles were used to design interesting wireless communication strategies (one involved sensor node timing offset correction based on firefly emission coordination, and the other involved sensor node sleep cycle scheduling based on positive/negative feedback in ant colonies). Along with genetic algorithms, such strategies tend to be quite neat and intuitively satisfying. The key challenge in this area is to prove why such strategies work well…or are such proofs even possible?