Research on UAV Sensor Attack Detection Based on Kalman Filter

Global Positioning System (GPS) sensor attack is a common attack that endangers the safe flight of (unmanned aerial vehicles) UAV. Based on Kalman filter, this paper studies the detection of UAV's GPS sensor attack and the security analysis of attack detection. Firstly, the dynamics and kinematics model of the UAV are established by using the method of flight mechanics, and the sensor attack signal is added to establish the GPS sensor attack signal model. Secondly, based on the principle of Kalman filter, the residual signal is generated, and the attack detection method of chi-square test is designed to detect GPS attack on UAV. Thirdly, the security of the attack detection method based on Kalman filter is analyzed, the stealthy GPS deception attack signal is designed to destroy the estimation performance of Kalman filter, and the stealthy attack detection strategy based on Euclidean distance is proposed. Finally, through simulation, the limitation of attack detection based on chi-square test and the effectiveness of attack detection based on Euclidean distance are verified.