Pan Li

A key step in ionosphere modeling is to convert slant total electron content (STEC) to vertical total electron content (VTEC) using an appropriate ionospheric mapping function (MF). This makes the accuracy of the applied MF a major factor affecting the performance of the built ionospheric model. Thus, this study evaluates ionospheric models constructed with three commonly used MFs, i.e., single-layer mapping function (SLM MF), modified single-layer mapping function (MSLM MF), and Klobuchar MF, from different aspects. Results show that the performance of the ionospheric model estimated with MSLM MF is the best, followed by SLM MF and Klobuchar MF. Compared with global ionosphere maps provided by the international Global Navigation Satellite System (IGSG), the MSLM MF-based ionospheric model achieves the best accuracy and consistency. During high solar activity, the SLM MF-based ionospheric model shows the largest discrepancies, which is up to about 40 TEC unit (TECU), while those of the MSLM MF-based model reduce to 10–15 TECU. Under low solar activity, the differences decrease, with the MSLM MF-based model maintaining the smallest bias and root mean square (RMS), which are -1.43 TECU and 2.48 TECU, respectively. The largest errors appear in equatorial regions (±15° latitude), where strong ionospheric gradients exist. Over oceanic regions, the MSLM MF-based ionospheric model still performs best, keeping bias smaller than 5 TECU and RMS smaller than 6 TECU, while the errors of the SLM MF-based ionospheric model can reach 15 TECU. Self-consistency assessments further confirm the stability of the MSLM MF-based ionospheric model, with bias and RMS below 1.0 TECU and 2.0 TECU, significantly outperforming those of the SLM MF-based and Klobuchar MF-based ionospheric models. Furthermore, the single-frequency precise point positioning tests also indicate that MSLM MF-based ionospheric corrections yield the best positioning accuracy, with horizontal and vertical errors typically ranging from 0.2–0.7 m and 0.4–1.0 m, respectively. Overall, the MSLM MF demonstrates the most reliable and consistent performance in ionospheric modeling under varied solar conditions, across diverse regions, and according to multiple metrics.

With the rapid development of Real-Time Precise Point Positioning (RT-PPP), the BeiDou-3 PPP-B2b and Galileo High Accuracy Service (HAS) have become pivotal for providing free, global, satellite-based precise correction services. However, most existing open-source decoding tools are developed based on the Python language, which limits their integration into embedded systems and real-time C/C++ positioning frameworks. In addition, these raw data streams lack standardized input formats and product generation, and do not take into account user-side satellite antenna calibration, which hinders offline analysis and cross-platform verification. To bridge these gaps, this study develops and presents the GKit-SSRDecoder, an open-source, high-performance C/C++-based software designed for the end-to-end processing of PPP-B2b and HAS messages. The software features a modular architecture that supports multi-receiver raw data stream parsing, State Space Representation (SSR) message decoding, and the automated generation of RINEX-compliant precise orbit, clock, and code bias products. Comprehensive evaluations were conducted using data from diverse GNSS receivers and compared against post-processed products. The results demonstrated that the Signal-In-Space Range Errors (SISRE) for the PPP-B2b service is 0.093 m and 0.086 m for GPS and BDS-3, respectively, while the HAS service achieves 0.057 m for GPS and 0.048 m for Galileo. Positioning tests showed that the generated products enable RT-PPP to reach centimeter-to-decimeter-level accuracy in static and kinematic modes. Specifically, in real-time kinematic scenarios, the positioning performance of the PPP-B2b and HAS services achieves RMS values of 0.13/0.11/0.12 m and 0.18/0.12/0.23 m in the East/North/Up directions, respectively. This software serves as a useful and portable tool for the GNSS community, facilitating the industrial application and scientific research of satellite-based augmentation services.

Loess landslides in northwestern China are highly unstable and difficult to distinguish due to sparse vegetation and their spectral and morphological similarity to the surrounding terrain. These landslides demonstrate considerable diversity in manifestation, encompassing shallow translational slides, small-scale features, partially obscured formations, and instances with irregular or poorly defined boundaries. To address the above issues, we propose MPES-YOLO, a multi-scale lightweight YOLO-based framework with selective edge enhancement to detect loess landslides. This model is based on the YOLOv8 architecture and incorporates a multi-scale partial convolution and exponential moving average (MPCE) module to improve multi-scale feature representation while reducing computational cost and enhancing small-target sensitivity. Additionally, to address ambiguous boundaries, a selective edge enhancement (SEE) module is introduced to extract authentic object edges from original images and inject them into key training layers, improving boundary perception. Finally, SIoU is adopted to improve geometric consistency for irregular landslide boundary localization. This paper first verified the basic detection performance of MPES-YOLO on the publicly available Bijie landslide dataset. Then, an experimental study was conducted in the loess landslides of Yan’an City, Shaanxi Province. The mAP@0.5 was 91.9%, and the parameter quantity was reduced by 23.3% compared with the baseline model. A generalization experiment was also carried out on the landslides in the Ningxia region, with the mAP@0.5 being 97.4%. The results show that MPES-YOLO achieves a strong balance between detection accuracy and computational efficiency, providing an effective and scalable solution for automated loess landslide detection and geological disaster early warning.

The advancement of Internet of Things (IoT) and autonomous driving applications has resulted in significantly increased demands for lane-level navigation using consumer-grade smartphones. However, the frequent occurrence of carrier phase cycle slips in smartphones severely degrades positioning accuracy, posing a major obstacle to achieving reliable lane-level navigation. While various cycle slip detection methods exist, their reliance on fixed and empirical thresholds often fails to adapt to the diverse characteristics of different smartphone brands, which can lead to either excessive false detections or missed cycle slips, thereby destabilizing the positioning solution. To address this gap, this study first quantifies the impact of small cycle slips on the sub-meter positioning accuracy of smartphones by inserting several simulated cycle slips in realistic dynamic driving scenarios. The results indicate that the impact of small cycle slips on lane-level navigation is at the centimeter-level, the reason might be that the small cycle slips are accommodated by the poor-precision smartphone phase measurements and ambiguity estimates. Consequently, only a negligible effect is brought on overall navigation performance, motivating the development of loose detection thresholds. Furthermore, using 211 vehicular dynamic datasets from the google smartphone decimeter challenge (GSDC), this study analyzes the distributions of the difference between Doppler and time-differenced carrier phase (D-TDCP) test values for different smartphone brands, including Xiaomi Mi8, Pixel, and Samsung. The analysis reveals that the distributions exhibit significant brand-specific differences and are strongly correlated with the signal-to-noise ratio (SNR). Based on these insights, an adaptive, SNR-fit threshold scheme for smartphone real-time kinematic (RTK) is proposed, dynamically adjusting detection thresholds for per smartphone brand and frequency. Validation on 103 independent test datasets demonstrates that, compared with conventional fixed and tight threshold methods, the proposed adaptive scheme reduces the 68th and 95th percentile horizontal positioning errors by 17 and 36 cm, respectively. Moreover, it increases the proportion of positioning errors within 1 m from 47% to 55%, inferring stable and reliable lane-level positioning capability.

With external ionospheric corrections, precise point positioning solutions can be augmented with proper constraints, enabling fast convergence and high-precision positioning performance. However, there are two main issues in user algorithms. Firstly, the contribution of the number of applied ionospheric corrections on positioning performance is not clear to users. For instance, in low-cost sensors where computation resources are limited, applying too many corrections (e.g., tens of corrections) from server will possibly cause slow response. Secondly, the determination of constraining strength for corrections from different satellites remains challenging. Therefore, this paper evaluates the effect of the number of ionospheric corrections on positioning errors and convergence time, and proposes an improved tight-constraining algorithm for partial reliable corrections. Static and simulated kinematic positioning results confirm that the more corrections applied, the better the positioning performance. However, the positioning accuracy only has limited improvements when several corrections are already applied, especially in the north direction. The proposed method reduces static positioning root-mean-square errors by 43.4%, 26.9%, and 32.1% in the east, north, and up directions, respectively, compared to solutions without constraint. In simulated kinematic results, the convergence time is further shortened by 7.7%, 4.8%, and 2.5% for the east, north, and up directions, respectively, compared to cases where all corrections are applied.

Smartphones, being one of the most ubiquitous sensors in daily life, have the capability to receive global navigation satellite system (GNSS) signals, thereby enabling them to provide location-based services (LBSs) for mass-market users. Spatial information is one of the vital components in intelligent transportation and Internet of Things applications, and transportation-related applications such as lane-level navigation are among the most frequently used smartphone LBSs. Considering the high noise level of smartphone GNSS measurements in such applications, there is a risk of selecting a reference satellite with measurement outliers in relative positioning technology, which would therefore decrease positioning accuracy and reliability. To address this issue, this article proposes a resilient reference satellite configuration in smartphone relative positioning, where two reference satellites are selected per frequency for each constellation, accompanied by an automatic switching strategy between single- and dual-reference satellite configurations. The proposed method extends the observation equations with a second reference satellite and is validated with 18 datasets collected in driving environments, and both theoretical analysis and positioning results demonstrate that the dual-reference satellite configuration outperforms conventional single-reference satellite strategies except in extremely harsh environments. When applying the resilient switch, the percentage of horizontal positioning errors within 4 m is largely improved. Moreover, the 68th percentile horizontal positioning errors are reduced by ~3 decimeters compared with single-reference satellite method, and the percentages of positioning errors within 1.0 and 1.5 m are improved by 8% and 9%, respectively, indicating a higher capability and great potential of providing lane-level navigation with the proposed resilient reference satellite configuration.