Dataset Construction for Anti-Spoofing Testing in BDS/5G/INS Integrated System

Existing public Global Navigation Satellite System (GNSS) spoofing datasets primarily support GPS L1 C/A and Galileo E1 OS signals, lacking BeiDou (BDS) signals, multi-sensor data, and realistic scenarios where spoofing, multipath, and platform dynamics coexist. This gap limits the testing of advanced anti-spoofing methods for multi-sensor integrated systems and emerging 'BDS +5G ' Position, Navigation, and Timing (PNT) architectures. To address this need, we construct multimodal datasets by developing a software-defined spoofing source for BDS B1I signals and designing a vehicle trajectory that incorporates platform maneuvers such as acceleration, deceleration, and turning. The datasets superimpose synchronous spoofing signals of multiple types and multipath interferences upon authentic signals at the digital intermediate frequency level. They also include raw inertial navigation system (INS) measurements inverted for three accuracy grades, as well as 5G angle and range observations derived from downlink signals in an Urban Microcell (UMi) scenario. Validation using a software receiver with maximum likelihood estimation (MLE) and an Extended Kalman Filter (EKF) confirms the effectiveness of the datasets in inducing intended timing and position spoofing with accurate multipath modelling. Under tactical-grade INS and interference-free conditions, the integration of 5G observations reduces the 3D positioning root mean square error (RMSE) from 1.57 m to 0.52 m, highlighting the complementary benefit of 5G. These datasets provide essential support for testing BDS and multi-sensor anti-spoofing algorithms under complex conditions.