An Approach of Linear Regression-Based UAV GPS Spoofing Detection

Table 1

Physical meaning of parameters in workflow chart of LR anti-spoofing model.

Parameters	Physical meaning
: LR predicts trajectory points	The continuous mission track points of UAV predicted by LR prediction model through historical track.
: GPS real-time data	At present, the UAV airborne sensors receive the real signal from the mission environment.
: corrected data	The value () transmitted to the UAV navigation system is selected according to the judgment of whether the current UAV mission environment is safe (whether there is GPS deception signal).
: noise threshold	From the analysis of flight experience, the reasonable path error of UAV in a safe and normal mission environment due to its own attitude control and physical environment is obtained.
: spoofing detection window with maximum length of	The model provides two detection means. The window is set for further detection of deception, recording the trajectory values of the UAV at five adjacent moments ( is set to 5 in the invention).
: error function	The variables involved in the calculation are LR predicted value and GPS real-time data at the current time. Compared with , the results are used to judge whether the current UAV mission environment is safe or not.