GNSS Spoofing Detection for Unmanned Aerial Vehicles
Submitted for my BE Mechatronics Hons Thesis in 2016 at The University of Sydney.
Imagine a world where we can secure systems by relying on their autonomous strengths.
The problem I saw…
In a world where adversarial teams can bring down UAVs with complex GNSS spoofing systems how can an autonomous system identify which signals to trust and which to ignore without encryption or mutual verification.
The opportunity sensed…
With signals being mimicked to an extremely high quality I decided to bank on the fundamental factors that differentiated a true satellite and a ground based spoofer (namely the altitude and sheer distance).
Knowing that the distances differed in orders of magnitude, the angles of arrival, phase shifts, and differentials in Time of arrival of signals compared to their first multipaths would present more disparately for ground based spoofers than satellites.
Beyond this, if the spoofer cannot predict the path being travelled by the UAV, a random selection from a series of preset manoeuvres could be made to incite the difference between spoofed and real signals.
Randomness and unpredictability can be a powerful tool in adversarial situations.
My solution…
Since we couldn’t transmit at these frequencies, I simulate with basic ray-tracing mathematics to test my hypothesis.
Under simplified constraints it proved that the theory presented a sound approach to solving the underlying problem in a manner that leveraged natural advantages.
