Listening to presentations at the recent Crowncom 2012 conference on “Cognitive Radio” gave some interesting insigths. On the positive note, there now seems to be a much better understanding among industry, regulators and academia on what is the role of secondary spectrum access in the future of wireless. Its not a simple picture – some systems are better suited for this spectrum regime (e.g. short-range indoor systems), others systems need primary status in licensed spectrum to work properly (e.g. cellular systems). This is also very much the findings of the QUASAR project.
On a more depressing note – although it now seems pretty obvious to most people in industry, at regulators and to others with insight in radio propagation characteristics, that secondary spectrum (“white space”) access to legacy system spectrum (e.g. TV spectrum) based on “cognitive sensing” simply doesn’t work. Still, there are lots of scientific papers on making even better sensing based access algorithms. If have the feeling that many take the highly idealized assumptions that are made in the (mostly excellent) work to get information theoretic bounds on the system performance for granted, and use those assumptions in proposing engineering solutions.
However, the Devil is in the details. There is, in fact, a (good) reason that the regulators have chosen the database approach for secondary access. With some back-of-an-evelope calculations its easy to see that even “perfect” sensing does neither on its own provide enough protection for primary users in most commercially interesting scenario, nor can it provide any useful information beyond the information one may get from a spectrum data base. There are a few exceptions, like Radars, but in these cases the primaries are so high powered that the actual sensing problem is more or less trivial. In his keynote talk, Petri Mähönen conjectured that the peak of academic papers on cognitive sensing base schemes may not yet have been reached its peak. He took the example of academic publication in the ATM-area in the late 90’s peaking after the technology was literary abandoned by industry. This is somewhat tragic… there are many other, really interesting problems in secondary spectrum access we need to address. There are definitely lots of applications for “cognitive” techniques (in Mitolas original sense) in wireless systems, but this particular one is not one of them.