Here's a paper that proposes GPS authentication based on signing and statistics about the signals (lots of math):
http://radionavlab.ae.utexas.edu/images/stories/files/papers...
Paper from Los Alamos with a receiver-only implementation that would use:
- abnormally high signal strength
- abnormally regular transmissions
- a secondary time source to double-check time (e.g. NTP on a smartphone)
- dead-reckoning based on accelerometers/gyroscopes/compass to double-check position
http://lewisperdue.com/DieByWire/GPS-Vulnerability-LosAlamos...