Slashdot Mirror
LightSquared Says GPS Tests Were Rigged
itwbennett writes "Would-be cellular carrier LightSquared claims that the company's LTE network was set up to fail in GPS interference tests. 'Makers of GPS (Global Positioning System) equipment put old and incomplete GPS receivers in the test so the results would show interference, under the cover of non-disclosure agreements that prevented the public and third parties from analyzing the process,' LightSquared executives said on a conference call with reporters Wednesday morning."
>old and incomplete GPS receivers
I'm not an expert in the deployment of GPS, but is this not what we would consider a real-world test? Why should they be set up to pass the test, by only testing the latest deployments of GPS?
Don't you test, in order to understand previous unknowns or to flesh out previously unforeseen scenarios?
---Up Up Down Down Left Right Left Right B A START
If this is fraud on the GPS companies' part or the testing authority's part then there should be hell to pay.
If this is sour grapes then LightSquared just libeled the companies involved.
If, on the other hand, "old and incomplete equipment" tests were a required part of the test for good reason, then LightSquared is a bit late in its complaints - it should've made these complaints well before testing happened, and its current statement should've started off with "As we said before the tests were run, testing for old and incomplete equipment is not a valid test...."
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
A lot of aircraft GPS receivers are quite old. It can cost 10-20K$ to put a certified receiver in a light aircraft, so pilots will keep their existing equipment as long as possible. Changing the requirements on interference resistance might require very expensive re-certifications of these receivers.
1) GPS manufacturers are not a direct competitor to a wireless networking company. If Verizon or AT&T were complaining they might have a case.
2) GPS was there first.
3) Clearly the Lightsquared hardware is spitting out a harmonic which could be fixed but would probably make the devices much more expensive to produce.
4) Lightsquared has been trying this case in the court of public opinion by running full page newspaper ads instead of dealing with the technology issues.
5) Lightsquared has been making huge political donations and receiving government grant funding which makes the whole thing stink like old fish.
it's not about filters, nor is it about "GPS listens outside its band"
GPS receivers have "wide open" front ends and always have for good engineering reasons:
1) Spectrum planning ensured that there's no high power signals in adjacent bands (i.e. the adjacent band is also for satellite signals)
2) "brick wall" filters are heavy, expensive, large, and have bad effects on the inband signals (see, e.g. any digital audio application since CDs started being sold 30 years ago). Your cellphone has GPS that is as small as it is partly because you can use a fairly wide open front end that doesn't require a lot of filtering.
3) GPS signals are below the noise floor, allowing use of 1 bit ADCs in receivers, reducing cost and complexity in receivers.
There's quite a bit of arguing about what is an appropriate propagation model from L2 terrestrial transmitter to GPS victim. L2 would like to use a conventional communication model. GPS folks would like to use a jammer/interference model. The difference isn't in the "mean power" but is in where the outliers are. For comm, your concern is that your worst case low power deviation is still high enough that you can "close the link" (i.e. not drop the call). For interference, your concern is that the worst case high power deviation is still low enough that it doesn't interfere with your link. The problem is that in urban environments, the distribution isn't uniform and is highly skewed (lots of reflecting surfaces and multipath.. distance isn't as big a factor as just the number of bounces). There's lots of deviations below the mean, but small ones, and relatively few deviations above the mean, but they are huge (e.g. "hot spots"). We're talking 15-20 dB difference between the 5% low end and the 5% high end
There's also arguing about what "performance degradation" is acceptable. L2 would like to claim that 6-8 dB is ok, while GPS industry would like to use 1dB. That's because communications uses error correcting codes and such, and can tolerate dropouts and degradation. GPS is more like radar, and relies on measuring the timing of the signal, and doesn't have as much in the way of error correction or error tolerance, so they've historically used the radar standard of 1dB degradation. The GPS industry is a bit stretching here, because with new receiver designs (which might consume more power and be bigger) they could probably deal with the worse interference environment. But that's a 10-20 year kind of project.
So the tests were fair, with published test criteria, and only now, a week from their deal with Sprint expiring (after a 30 day reprieve) they're starting to raise these questions.