We are planning the design of an anti-jamming systems for UAV and presented it to the Galileo Master 2012 competition. The idea comes analyzing the recent case of the UAV landed by jamming its navigation system. By putting noise [jamming] on the communications with the satellite (the UAV was far from its closest control station) , they probably forced the UAV into autopilot RTL mode (Return To Launch).At that time the UAV relies on GPS signals to get home. By spoofing the GPS, jammers were able to get the UAV to 'land on its own where they wanted it to, without having to crack the remote-control signals and/or satellite communications data link.'"
The idea is actually nothing special, it uses redundancy, i.e. uses more independent GPS sensors controlling each other on the same platform and then let just one of them being 'fooled' by the jamming, while the other maintain correct information and drive the UAV back home.
A compass and some accelerometers (or even a view of the sun and an RTC) are a lousy substitute for the accuracy of GPS; but they do provide a sanity check that could keep you going in approximately the right direction, at least enough to hard-land somewhere nominally friendly, if GPS cannot be trusted.
It's certain that the UAV had an inertial navigation system - the problem is, how do you know when to use it? The way they usually work is that the navigation system computes two solutions: a hybrid GPS/INS solution to use most of the time, and a backup inertial-only solution. The inertial-only solution doesn't get used by the flight computers unless GPS is out entirely or there's some other very obvious problem. If you spoofed a GPS signal with real coordinates and slowly guided it away, how could the nav system see there's something wrong?Inertial navigation systems need reference points to prevent huge drifts over time. This is especially a problem if the aircraft flies relatively straight at the same speed for a long time -- accelerometers won't be able to detect slight changes in course. GPS is used to provide the reference points to reduce the drift. If the GPS system is wrong, then the inertial navigation system is also going to be fooled.
RoadmapWe first will test a solution with legacy Ardupilot with Air Shield (pressure sensor), no thermopiles, and an ArduIMU V2 instead, connected to a Mediatek 10Hz GPS. This with be 'sacrificed' and will talk to the jamming system (jammer), trying to make sure he thinks he is in control and following its instructions.Another two GPS receivers, i.e. a uBlock (binary) and and EM-406 Sirf with different protocols would stay behind, protected from the jammer by a firewall, and fly the UAV back to a safe home position. The system is based on open-source data and limited, through Xbee telemetry, to 1.2 km. Next step will be performed on a remote system up to 40 km and finally on a satellite datalink.