The traditional triple-frequency geometry-free pseudorange minus phase (GFPMP) method is very susceptible to the influence of pseudorange observation noise, and it is difficult to detect insensitive small cycle slips. The dual-frequency phase ionospheric residual (PIR) method has the ability to detect sensitive small cycle slips, but its detection results have multivalue problems. In view of this, a GFPMP-PIR combination model method with wavelet denoising (GPW) is proposed here. The idea of this method is as follows: firstly, wavelet transform is used to denoise the pseudorange observations; then, by optimizing and selecting the combination coefficients with the minimum condition, an optimized GFPMP-PIR combination model is constructed. Hence, the source data quality is assured through wavelet denoising. And the accuracy of cycle slip detection and repair is improved by using the dual-frequency PIR combination. Moreover, the multivalue problem of the dual-frequency PIR method is solved by triple-frequency GFPMP combination; finally, the BeiDou navigation satellite system (BDS) triple-frequency measurement data is used for experimental verification. The experiment results show that the GFPMP-PIR optimization combination model with wavelet denoising can detect and repair various cycle slips, especially for insensitive small cycle slips.
To address the performance degradation of Inertial Navigation Systems (INS) during Global Navigation Satellite System (GNSS) signal outages in complex environments, this paper proposes an error compensation method for INS Non-Holonomic Constraints (NHC) based on Gated Recurrent Units (GRU). The method leverages the GRU neural network to generate pseudo-measurements during GNSS signal outage. These pseudo-measurements are then fused with INS NHC information using an optimized Adaptive Kalman Filter (AKF), forming an error compensation model for the integrated navigation system. Additionally, a training optimization strategy based on the ratio test value is introduced, which dynamically adjusts the weights of samples in the loss function, enabling differential treatment of measurements with varying confidence levels. The effectiveness of the proposed method is evaluated across three representative trajectory scenarios. Experimental results demonstrate that, compared to traditional methods, the proposed approach reduces positioning errors by more than 25% in all three scenarios, confirming its superior performance and robustness during GNSS signal outages.
19 September 2025 in Earth sciences and subsoil use