[time-nuts] Time security query

[Lux, Jim (337C)]

> 
> Spoofing a GPS receiver should not be too hard.  I would record the GPS
> spectrum off the air, then play it back, delayed by some sufficient time to confuse
> whatever you are trying to confuse.  By playing back actual signals, the GPS receiver would
> hear a self consistent set of signals, just shifted in time.  You would have to be close
> enough  to the target GPS,
> so that your spoofing signal was much stronger than the off the air
> signals.

Not too hard -> trivially easy.  A L band antenna, an amplifier, and another antenna with enough isolation so you don't make an oscillator will work.  The coax in the system provides the needed delay.  This is the origin of the "GPS jammer from radio shack" stories.

Note that the vulnerable time is during code acquisition.  Once the receiver is tracking a signal, another signal at more than 1 chip delay won't show up in the correlator output, unless it's MUCH (10s of dB) stronger.  However, during acquisition, the search logic tends to pick the strongest signal to lock on to. Since satellites are always appearing and disappearing, the code acquisition process occurs fairly often, although more sophisticated receivers are better at rejecting inconsistent data (e.g. they could reduce the search space when a signal fades, on the assumption that when that signal comes back, it will be in pretty much the same place)