As users increasingly prefer mobile devices for accessing the Internet, wireless networks that progressively exploit diverse unlicensed spectrum blocks are becoming an attractive alternative to cellular systems. This talk addresses the efficient evaluation and selection of the vast association options exposed by such wide-spectrum networks, where scanning-based or historical data-driven approaches become prohibitive. The talk will introduce a technique that infers channel quality and spectral usage from limited measurements collected in other bands and at previous locations, in order to select resources that optimise the performance objectives of mobile clients. Experimental results from a wide-spectrum testbed, spanning a spectral range of 5.085 GHz, will illustrate that this solution achieves significant throughput gains over fixed-band prioritisation and scanning-based practices.
Paul Patras is a postdoctoral research fellow at the Hamilton Institute, Ireland. He obtained a Ph.D. in telematics engineering from University Carlos III of Madrid in 2011 and is the first alumnus of the Institute IMDEA Networks, Spain. In 2010, he was a visiting researcher in the Networks Group at Rice University, USA. He has served on the program committees of several conferences and workshops, including IEEE VTC, IEEE WoWMoM and IFIP Networking, and was awarded the IEEE Communications Letters Exemplary Reviewer distinction in 2011. His research interests include performance optimisation of wireless networks, dynamic spectrum access and prototype implementations, and he is currently involved in the EU funded ICT FLAVIA research project.