基于多传感器数据融合的无人机GPS欺骗检测研究

GPS Spoofing Detection of UAV Based on Multi-sensor Data Fusion

摘要:

本文针对实际过程中可能存在的无人机GPS欺骗情况，提出了基于多传感器数据融合的GPS欺骗检测方法。该方法通过比较多传感器惯性导航系统加上Elman神经网络修正得到的位置信息与GPS输出位置信息，从而判断无人机GPS是否受到欺骗。该方法有两个创新点，第一个是使用Elman神经网络，在不增加传感器成本的基础上其有助于提高惯导系统输出位置信息的精度；第二个创新点是使用带延迟的扩展卡尔曼滤波器，用于解决多传感器数据不同步的问题。实验结果表明，本文提出的方法能有效的检测出GPS欺骗，从而保证无人机的安全飞行。

Abstract:

This paper proposes a GPS spoofing detection method based on multi-sensor data fusion for the possible GPS spoofing of UAVs in the actual process. The method compares the position information corrected by multi-sensor inertial navigation system and Elman neural network with the position information of GPS output, so as to determine whether the UAV GPS is deceived. The method has two innovations. The first one is to use Elman neural network, which is helpful to improve the accuracy of the output position information of the inertial navigation system without increasing the cost of sensors. The second innovation is to use the extended Kalman filters with delays for solving the problem of multi-sensor data out of sync. Experiments show that the proposed method can effectively detect GPS spoofing, thus ensuring the safe flight of drones.