Consumer drones support fail-safe mode to handle emergency situations such as a crash, low battery, or loss of the remote control signal. In these situations, drones force their flight mode to fail-safe mode. For example, the fail-safe mode forces the drone to return to the start position when the battery runs low. In addition, the fail-safe mode activates an emergency stop when the drone detects a crash to reduce damage to the drone body and people near the drone.
Severe GPS errors are also considered an emergency situation and are handled through the fail-safe mode in consumer drones. Basically, MEMS sensors, mainly used in consumer drones as inertial measurement units (IMUs), are subject to sources or error such as misalignment and temperature change. In reality, most consumer drones do not allow to activate autopilot without GPS. Thus, GPS fail-safe flight mode of consumer drones should be either the positioning mode which does not utilize GPS or landing mode because it is too dangerous to maintain autopilot based on only IMUs without GPS. In the case of drones whose GPS fail-safe mode is the positioning mode, there exists three different ways after the drones lock onto GPS signal again according to our analysis, and we can classify consumer drones according to these behavior types. Including the case of drones whose GPS fail-safe mode is landing mode, our classification scheme is summarized in the following table.
A brief GPS error can occur when GPS spoofing begins, so safe-hijacking strategies should be designed considering fail-safe type of the target drone. The corresponding safe-hijacking strategy for each drone type is summarized as follows.
Type I drones are trying to stay over their original position. If the attacker spoofs the target drone’s GPS position as if the drone is moving in a certain direction, drones are considered to drift owing to external factors such as wind. Thus, the target drone generates speed in the opposite direction, so the drone moves in that direction in the real world.
The strategy against type II drones is based on the drone’s characteristic that control their body according to their path-following algorithm, which enables them to reach their destination reliably and accurately during autopilot. If the GPS position is manipulated as the drone deviates from the path, it will move in a different direction from the original direction to return to the track. The moving direction is determined by their path following algorithm and the fake position.
In the strategy against type III and IV drones, the focus is to prevent the target drone from losing GPS lock despite GPS signal manipulations. This strategy is composed of two steps: 1) conduct soft GPS spoofing to make the target seamlessly lock onto the spoofed signal and 2) move the spoofed location adequately according to the target drone's path following algorithm and how it detects GPS error.