I do hope cognitive radio designs will be done responsibly for the spectrum they occupy, and I cite RMS Express as an example of a responsible approach to mitigating interference. And (military) ALE as I've experienced it as the opposite.
However, I fear device manufacturers wanting to use spectrum everywhere will not produce radios able to detect weak emissions when their receiver bandwidth is so wide as not to see it above the noise. Among the BPL comments and replies is one manufacturer's assertion that there were no signals to be interfered with -- when his spectrum analyzer noise floor was higher than the level those signals would normally reach. By using only measurement technology to required for Part 15 certification, that manufacturer was able to ignore signals I believe he "knew or should have known" (as the lawyers say) were or could be present. We must listen first. So should any responsible user of shared spectrum. He must be able to hear *any users authorized* in the spectrum shared, at levels and in bandwidths they are authorized to use. This is not so easy, considering that we often carry on Olivia or Contestia QSOs below the background noise level. It could be made easier by restricting automatic (cognitive) radio to spectrum where weak signal modes will not be encountered. Cortland KA5S > [Original Message] > From: Bob McGwier <[email protected]> > To: <[email protected]> > Date: 12/16/2009 12:54:35 AM > Subject: Re: [digitalradio] cognitive radio systems;? > > Cortland Richmond wrote: > > > > > > One problem with "cognitive radio" is that it seems it will be designed > > to detect only emissions similar to those it is meant to receive. > > Therefore, it is best used in spectrum particularly allotted to > > just those kinds of emissions. This rather defeats the purpose of > > "white space." > > > > RMS Express by way of contrast has a busy detector that will prevent > > transmitting over many kinds of modulation different than it uses. > > Compare this with (say) ALE, whose polling (encountered on MARS > > frequencies) takes no account of voice or even Olivia on channels it > > happens to select. > > > > > > Cortland > > KA5S > > > > > > This is not correct in my experience. In all serious systems under > development, the CR is looking to characterize all energy to some degree > or another, irrespective of whether it is a "matched filter" to a > particular waveform. > > The purpose is to find a channel that works. Energy on the channel is > an indicator it would not as the source would be cochannel interference > and with some high degree of probability, the interference would be mutual. > > Dislike for any particular system which automates channel usage but does > not behave responsibly is not to be used to condemn responsible digital > system developers. The enforcement of this responsibility is done by > pressure (peer) and performance (being interfered with by those not > detected). > > Bob > N4HY
