Elephants, Polo Ponies, Cognitive Radio & LTE+
A friend of mine showed me his copy of the The International Encyclopedia of Shooting today. Ok this does sound like it could be the start of a Mrs. Moneypenny article in the FT Magazine but any acronyms will be telecom acronyms. The following is a quote from the book relating to advice on how to shoot an elephant -
`When the elephant is down it is advisable to run up and put another bullet into the back of the neck, for a beast that is stunned can be up in an instant and move as fast as a polo pony‘
I actually photographed the page least you think I am making this up.
Anyway the polo pony is the reference point. The benchmark. A way of understanding fast. I thought of this again later on in the day when we were having a meeting about cognitive radio and LTE-Advanced.
It remains the case that cognitive radio has no polo pony. It has no benchmark. We discuss this regularly and in one of our FP7 projects we are in fact working on benchmarking but the reality is there is no way of judging how much faster or better the cognitive radio is than the polo pony because there is no polo pony.
HOWEVER – I think LTE-Advanced could be the polo pony we need.
LTE-Advanced is a cognitive network waiting to happen (currently minus unlicensed and dynamic spectrum access options but near enough).
When explaining cognitive radio I refer to the meters and knobs definition a lot. It is obvious that LTE-A has added a significant number of knobs to the radio and the network. There are multiple antenna configurations, multiple combinations of aggregated spectrum, multiple ways of load balancing across primary and secondary cells, multiple ways of scheduling the resources over space, frequencyand time, multiple topology options, multiple frequency planning and reuse factor combinations, multiple ways of managing interference, multiple uplink and downlink configurations, multiple RF frontend options, multiple RF chain combinations, multiple CFO techniques etc… and when you think of all the possible combinations and permutations of all of these variables, the mind boggles. And this is before dynamic spectrum access and spectrum sharing techniques are added (though not in any spec it is not difficult to imagine these additions). The reality is however, to get the performance we need out of the system the correct combinations of these options need to be exploited. In my opinion the only way to make this work is to think cognitive radio – to deploy and/or add all possible meters to understand and gauge conditions and context and to deploy advanced autonomous decision-making processes to determine the actions (configurations) that are needed.
This is what I mean by a cognitive radio waiting to happen.
So back to polo ponies. LTE-A should be the polo pony of cognitive radio. If my cognitive radio is more effective, more capable of self-organisation and self-configuration, cheaper, more power efficient, then it outruns the polo pony. That is what we need to do to progress the field and really move on. Make LTE-A a showcase for cognitive radio and then move beyond that.