| Abstract: | Global Navigation Satellite Systems (GNSS) have become an indispensable core component of modern automatic carrier landing systems, providing high-precision carrier phase observations. However, in complex wartime electromagnetic environments, carrier-based platforms often employ anti-jamming antennas using Space-Time Adaptive Processing (STAP) to suppress strong interference. The dynamic beam switching and weight updating in this non-linear signal processing introduce phase biases, causing impulsive and step-wise jumps in double-differenced (DD) carrier phase observations. If these non-physical phase distortions are not promptly identified, they severely threaten the integrity and availability of the approach and landing system. Existing fault detection methods face significant limitations when addressing these anomalies. The direct Triple-Differenced (TD) method suffers from a "dual-peak effect," making it difficult to determine the exact start and end points of a fault. Conversely, methods relying on the standard Kalman Filter (KF) innovation sequence suffer from "innovation decay" due to state pollution; the anomalous deviations are absorbed into the state estimation, rendering the standard KF unable to continuously detect long-lasting step-wise jumps. Furthermore, heterogeneous observation noise across different multi-receiver baselines makes fixed global thresholds ineffective. To overcome these challenges, this paper proposes a novel double-differenced carrier phase jump detection algorithm based on a Frozen Kalman Filter (FrKF) with weighted sliding window statistics. The novelty and technical contributions are as follows: ? Firstly, a fault isolation strategy based on a frozen update mechanism is proposed. When a jump is detected, the filter enters a frozen mode that cuts off the fault data from polluting the state estimation. This maintains the significance of the detection statistic and enables continuous locking and tracking of step-wise fault intervals. ? Secondly, a Kalman Triple Difference (KTD) statistic and an adaptive threshold model are constructed. The KTD statistic accurately indicates the fault epoch with a single pulse, resolving the dual-peak ambiguity of the traditional TD method. Simultaneously, a weighted sliding window is designed to dynamically track noise characteristics, which adapts to the heterogeneous noise of different baselines and maintains high sensitivity while significantly reducing the false alarm rate. ? Thirdly, the proposed algorithm is extensively evaluated using real datasets collected from a high-precision integrity monitoring platform equipped with multi-receiver anti-jamming array antennas. Preliminary results demonstrate that the FrKF algorithm provides superior robustness against both impulsive and step-wise fault modes. Unlike the traditional TD method which cannot track continuous faults, and the standard KF method which suffers from early alarm disappearance, the FrKF algorithm achieves 100% time-domain coverage for step-wise faults until the fault disappears or the system resets. It successfully isolates phase distortions in real-time without interrupting navigation solutions. |
| Published in: |
Proceedings of the 39th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2026) September 14 - 18, 2026 Hyatt Regency Grand Cypress Orlando, Florida |
| Pages: | 646 - 657 |
| Cite this article: | Zhang, Wentai, Wang, Yu, Sun, Chao, Feng, Wenquan, Zhang, Tianshuo, "A Frozen-Kalman-Filter Method with Fixed Thresholds for Double-Differenced Carrier-Phase Jump Detection," Proceedings of the 39th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2026), Orlando, Florida, September 2026, pp. 646-657. https://doi.org/10.33012/2026.20788 |
| Full Paper: |
Sign in to download with your ION membership, download credits or subscription, or buy this article. Sign In Buy this article: $12 for ION members, $35 for non-members |
| References: |
Show 19 references
|