GPS Spoofing Detection and Mitigation for Drones Using Distributed Radar Tracking and Fusion

Global positioning systems (GPS) receivers provide real-time position, navigation, and timing (PNT) services and have broad applications in navigation, aviation, maritime, transportation, and UAVs. The GPS blueprints are readily available, making generating and duplicating a GPS-like signal easy. It may lead to attacks and disruption, like jamming and spoofing, where corrupt GPS-like signals are projected onto the receiver to mislead the PNT of the GPS receiver.

These attacks endanger navigation, especially in UAVs or drones. The GPS receivers are widely used on unmanned aerial vehicles (UAVs) for localization, navigation, tracking, mapping, and timing. UAVs with GPS are subject to deliberate interference and spoofing attacks. In the spoofing process, the GPS's actual and predicted locations disagree. As a result, the UAV location predicted by GPS differs from the actual location, causing the application to fail.

This paper proposes GPS spoofing detection and mitigation for UAVs with a novel methodology of distributed radar tracking and established communication. In the proposed method, distributed radar ground stations are equipped with a local tracker to detect and eliminate the time-varying UAV dynamics. Moreover, in the proposed approach, UAVs and local trackers are connected to the fusion node (FC) to exchange information through the existing communication channel.

The position and covariance of the UAVs are estimated using the extended Kalman filter framework, which is sent to a fusion node as primary data. Simultaneously, UAVs time-varying kinematics are estimated using the extended Kalman filter and global nearest neighbor association tracker frameworks, and data is transmitted to the central fusion node as secondary data. The FC is responsible for determining the spoofing effect and mitigating the UAVs by transmitting the information.

At FC, a track-to-track association is performed to detect spoofing attacks on UAVs by using available primary and secondary data. The radar sensor acts as an additional source of information. Using a reference position from another sensor is beneficial since it does not require any alteration in the existing GPS hardware and signal processing techniques and can easily be adapted to the existing GPS receivers. The proposed track-to-track association mathematical formulation for the spoofing detection method can be employed for other sensors like visual, lidar, range, IMU, etc.

After detecting the spoofing attack, the secondary data is subjected to a correlation-free fusion algorithm. In the established communication channel, this fused state information is provided to the UAV as a control input to mitigate the spoofing attack. There is no restriction on establishing communication; any communication like Lora, 5G, or IoT module serves the purpose.

The spoofing scenario results show that using the fusion state from the established communication channel provides the same accuracy as a GPS receiver in a clean environment. Since the innovation is calculated using the predicted fused state, there is no effect on the number of satellite signals on position root mean square error (PRMSE). Regarding PRMSE, radars with low measurement noise outperform radars with high measurement noise. The proposed radar tracking and communication algorithm best suits drone swarm applications.