Addressing the challenge of extracting coseismic offset signals from GNSS coordinate time series using earthquake catalog information, this study introduces a method based on optimized window and component selection in Singular Spectrum Analysis (SSA). The proposed approach automatically selects relevant components for time series reconstruction according to the magnitude of eigenvalues and the contribution rate of each component, without relying on prior information or assumptions. This effectively overcomes the limitations associated with arbitrary decomposition and reconstruction of time series in conventional SSA. By subtracting the reconstructed series from the original data, a residual time series is obtained, from which reliable offset signals can be extracted as comprehensively as possible. Through simulation tests and an application to the March 9, 2011, Mw7.3 earthquake event in Japan, comparative experiments were conducted to evaluate different window lengths and varying numbers of reconstructed components. The results indicate that a window length of 365 days, combined with the selection of principal components whose cumulative contribution rate exceeds 95%, constitutes the most suitable strategy for detecting offset signals in GNSS residual time series.
Fault detection and identification algorithm is essential to ensure the reliability of global navigation satellite systems (GNSS) systems for high-precision intelligent applications. Low-cost receivers always suffers from GNSS pseudorange gross-error detection and carrier-phase cycle-slip anomaly in complex environments, which would degrade positioning accuracy. To address these issues, this paper proposes an enhanced low-cost RTK precise-positioning algorithm with an adaptive strategy, designed to detect multiple pseudorange gross errors and carrier-phase cycle slips. By leveraging time differenced darrier phase (TDCP) observations, the proposed method dynamically adjusts detection thresholds in real time, thereby minoring multiple faults and improving positining accuracy. To evaluate the performance of the proposed algorithm, a dynamic vehicle experiment equipped with ublox and BeiYun low-cost GNSS devices was conducted. Four schemes including GL, GLF, GLFTC, and GLFTA methods are compared. Results show that the 3D RMS positioning errors of the proposed GLFTA method achieves maximum 19.34% improvement. That is because the proposed GLFTA method incorporating an adaptive strategy can detect small carrier phase cycle slips than the GL, GLF, and GLFTC methods, thereby enhancing RTK positioning performance in complex environments.
Our study employs GPS coordinate time series and the NIF method to invert postseismic deformation, investigate the spatiotemporal evolution of fault slip, and evaluate the inversion efficiency of the NIF method under various parameter settings, thereby providing prior information for future research on earthquakes and fault slip. Taking the GPS coordinate time series from the Tokai region of Japan during 2009-2021 as an example, we conducted an analysis using the NIF method. The results indicate that the fault slip following the 2011 earthquake was primarily concentrated in the northeastern part of Tokai region (34.8°N-35.8°N, 137.2°E-138.4°E), with a maximum cumulative slip of 9.31 cm at depths ranging from 20 to 60 km. In terms of model parameter analysis, the inversion time and memory cost increase with the number of sub-faults, while the maximum cumulative slip approaches a stable value. Variations in the proportion of local benchmark motion proportion parameter do not affect the inversion time or memory cost but are negatively correlated with the maximum cumulative slip. The time scale of the GPS coordinate time series has no significant impact on the NIF inversion. In contrast, higher time resolution of the GPS coordinate time series results in longer inversion times and greater memory cost, with the maximum cumulative slip gradually stabilizing.
Addressing the inaccuracies in the ground-calibrated phase center offsets (PCOs) provided by the China Satellite Navigation Office (CSNO) for BeiDou-3 (BDS-3) satellites and the absence of phase center variations (PCVs), this study performs in-orbit calibration of the B1C/B2a antenna phase centers for BDS-3 MEO and IGSO satellites under the IGS20 frame. Using ionosphere-free combination observations from 155 global IGS/MGEX stations spanning 2021 to 2024, we demonstrate high consistency in PCV estimates for satellites of the same type, with most nadir-angle PCV differences below 1 mm and an overall standard deviation of approximately 0.3 mm. Horizontal PCO estimates align with igs20.atx values at the centimeter level. After aligning the Z-PCO estimates with the IGS20 frame, the mean differences with respect to the igs20.atx values are -9.5 cm for MEO satellites and 40 cm for IGSO satellites. Compared to the igs20.atx model, the phase center correction (PCC) model derived in this study improves the orbit overlap accuracy for BDS-3 MEO satellites by 0.43% to 7.03%. Furthermore, using the estimated B1C/B2a PCC model, the mean scale factor difference for BDS-3 solutions relative to the IGS20 frame is 0.16 ppb, confirming successful alignment with IGS20. The BDS-3 PCC results have been submitted to the International GNSS Service (IGS) through the Chinese Academy of Sciences (CAS) Analysis Center, contributing to integrated analysis and promoting the high-precision international application of the BeiDou Navigation Satellite System.
With the arrival of the maximum phase of Solar Cycle 25, ionospheric scintillation phenomena have become increasingly frequent, which severely impacts the capabilities of GNSS positioning, navigation, and timing services. This study collected 30-second sampling rate data from 6 GNSS stations in low-latitude regions of Asia, Oceania, the Pacific, Africa, and South America spanning the period 2014-2024. Using S4c as the criterion for identifying low-latitude amplitude ionospheric scintillation, the statistical characteristics of ionospheric scintillation in various regions were analyzed. The results demonstrated a significant positive correlation between the occurrence of ionospheric scintillation in various regions and solar activity. During solar maximum years, ionospheric scintillations at the KOUG station mainly occurred in winter, spring, and autumn, while those at the other stations are predominantly observed in spring and autumn annually. Scintillation at all stations predominantly takes place between 8:00 p.m. and 2:00 a.m. local time. In low solar activity years, both the frequency and duration of ionospheric scintillation at each station decrease significantly. With the exception of the KOUG station, where scintillation remained severe, the other stations show no obvious seasonal or local time patterns. The directional distribution of ionospheric scintillation exhibits a distinct relationship with the latitude of the stations. Furthermore, the study indicates that due to the distinct meridional distribution characteristics and latitude-dependent variations of equatorial plasma bubbles, the scintillation intensity at different locations during high solar activity years follows the order from strongest to weakest: South America (near the Atlantic Ocean), Africa, the Pacific region adjacent to South America, Asia, and Oceania.
The El Niño-Southern Oscillation (ENSO) is a complex ocean-atmosphere interaction phenomenon that drives extreme weather events globally, with water vapor playing a crucial role in its evolution. Recently, Global Navigation Satellite Systems (GNSSs) have emerged as an effective tool for retrieving water vapor with high accuracy, high spatial and temporal resolution, and all-weather capability. However, GNSS-derived precipitable water vapor (PWV) has not been well researched for its potential in the study of ENSO, particularly regarding their lead-lag relationships. This study investigates the spatiotemporal response relationship between PWV variations and the Oceanic Niño Index (ONI) using 12 years of coastal GNSS-derived PWVdata. Multichannel singular spectrum analysis (MSSA) was employed to extract nonlinear trends of PWV anomalies, followed by Pearson correlation and lead-lag correlation analyses with ONI. The results reveal a moderate negative correlation in the western Pacific and positive correlations in the eastern Pacific and western Indian Ocean. Notably, stations closer to the Niño 3.4 region exhibited stronger correlations. Moreover, more than half of the stations showed absolute correlation coefficients exceeding 0.4 at optimal lag times, indicating that ENSO exerts a lagged influence on PWV at most stations. A case study at the COCO station near Indonesia (2015-2017) demonstrated that PWV and precipitation anomalies lagged ONI by approximately 8 months, consistent with the severe drought in 2015 and flooding in 2016. These findings suggest that GNSS-derived PWV serves as a valuable indicator for monitoring ENSO dynamics and could enhance early warning systems for ENSO-related drought and flood risks.
To further enhance the accuracy of atmospheric weighted mean temperature (Tm) models in ground-based Global Navigation Satellite System (GNSS) retrieval of precipitable water vapor (PWV), we propose and develop a regionally adaptive Multi-Hidden-Layer neural network for Tm, hereafter referred to as MHL_Tm. A multiparameter cooperative Tm-modeling framework has been established using radiosonde observations from 65 launch sites across China during 2014–2018. We analyzed the nonlinear coupling between surface temperature (Ts), surface water-vapor pressure (e), latitude (Lat), elevation (H), and the temporal factor day of year (DOY) with radiosonde-derived integral Tm values. Radiosonde data from 2019 served as an independent reference to evaluate MHL_Tm’s performance, which was then compared against the Bevis, GPT3, and Elastic Net models. Experimental results showed that the annual mean bias of MHL_Tm was –0.61 K, representing reductions of 30 % and 58 % relative to Bevis and GPT3, respectively, and slightly higher than Elastic Net (0.11 K). The annual mean RMSE of MHL_Tm was 2.77 K, corresponding to improvements of 35 %, 62 %, and 18 % over Bevis, GPT3, and Elastic Net, respectively. Across different latitudinal and altitudinal zones in China, MHL_Tm exhibited superior accuracy and stability compared to Bevis, GPT3, and Elastic Net, demonstrating excellent regional applicability.
To address the challenges of energy constraints, high real-time requirements, and strong adversarial conditions in underwater dynamic target pursuit tasks, a highly efficient two-layer dynamic target pursuit algorithm is proposed to overcome the limitations of multi-autonomous underwater vehicle systems in terms of coordination efficiency and task execution. First, by analyzing the relative velocities of pursuers and the target, the Apollonius circle principle is extended to three-dimensional space, facilitating a pursuit strategy that aligns more effectively with real-world underwater conditions. Second, to mitigate the inherent measurement errors of sonar detection systems, an adaptive Kalman filter is designed to effectively suppress noise interference and achieve real-time, accurate prediction of the target AUV's motion trajectory. Furthermore, by reconstructing the neuronal activity propagation mechanism of the Glasius bio-inspired neural network, the collaborative decision-making process of multiple AUVs is optimized, significantly enhancing task execution efficiency. Simulation results demonstrate that the proposed algorithm improves pursuit distance and time by at least 30% and 24%, respectively. In multi-scenario generalization tests, the average pursuit distance and time are improved by at least 25% and 18%, respectively. In anti-interference tests under varying sonar detection accuracies, the average pursuit distance and time are enhanced by at least 25% and 22%, respectively. These results collectively validate the superior accuracy, robustness, and adaptability of the proposed algorithm.
This manuscript is focused on standardizing the process of the a posteriori precision evaluation in discrete Kalman filtering. Although the a posteriori precision evaluation of the solution was considered as indispensable within the method of least squares, the solution of a Kalman filter shows a lack of a posteriori precision evaluation for too long. Even worse, there often exists appalling confusion about what is considered as the a posteriori precision of the solution in Kalman filtering. The authors hereto propose to put the a posteriori precision evaluation of the solution into practice at four different levels in Discrete Kalman filtering through estimating: (i) the a posteriori variance of unit weight (or reference variance), (ii) the separate a posteriori variance factors for the process and measurement noise vectors, respectively, (iii) the individual a posteriori variance factors for the independent noise groups, and (iv) the individual a posteriori variance factors (or components) for the independent process noise factors and measurement types. A working example is presented to illustrate the proposed a posteriori precision evaluation in Kalman filtering using a road test based on the double-differenced GPS L1 C/A, L1 and L2 carrier phases and the specific force and angular rate measurements from an MEMS IMU. With the rapidly increasing utilization of the Kalman filter in modern applications, the inclusion of the proposed a posteriori solution precision evaluation, as part of the standard solution, in discrete Kalman filtering is not only necessary, but also can be expected to happen soon within our grasp.
This paper presents a comprehensive evaluation of the Mosaic-H GNSS receiver's RealTime Kinematic (RTK) performance when integrated with unmanned aerial vehicle (UAV) platforms. Through systematic testing across varied operational environments including open-sky, urban canyon, and forested areas, we characterize the receiver's positioning accuracy, fix reliability, and multipath resilience. Our experimental methodology employed a precisely configured base station broadcasting RTCM corrections via Port 8105, with the UAV-mounted receiver maintaining centimeter-level accuracy in optimal conditions. Results demonstrate an 88.2% RTK fixed solution rate across all test scenarios, with horizontal accuracy averaging 3.1 cm in open environments. Notable findings include the identification of critical operational thresholds: maintaining >10m clearance from buildings prevents multipath-induced degradation, canopy coverage exceeding 70% triggers fallback to float solutions, and vehicle speeds above 8 m/s challenge phase tracking capabilities at 5 Hz update rates. The multipath impact score correlation analysis revealed strong relationships between environmental factors and positioning quality (R²=0.87), enabling predictive mitigation strategies. These findings establish operational guidelines for reliable centimeter-accuracy UAV navigation in complex environments, with direct applications to precision agriculture, infrastructure inspection, and autonomous aerial surveying.
Accurately assessing the hazards posed by landslides is of great importance for disaster prevention and mitigation. This study proposes a method of landslide hazard levels analysis based on displacement traction, a term referring to the correlated directional influence between surface displacement vectors at GNSS (Global Navigation Satellite System) monitoring points. By analyzing these spatial correlations, the optimal grid unit size is determined for refined hazard levels assessment. To construct a representative target area, the improved Sparrow Search Algorithm was combined with the k-means clustering algorithm, integrating displacement characteristics and the derived grid unit size. A hazard assessment dataset was then developed for the target area. Subsequently, a stacking ensemble model was employed to evaluate landslide hazard levels using eleven influencing factors, including surface roughness, elevation, and slope angle. Experimental results demonstrated that the proposed method outperformed the conventional fixed-grid approach, yielding a 2.15% improvement in overall accuracy, a 1.30% increase in F1-score, and a 3.75% gain in the kappa coefficient. This study not only enriches the theoretical foundation of assessing landslide hazard levels but also provides a powerful technical support and practical guidance for the scientific prevention and control of landslide disasters.
In the Global Navigation Satellite System (GNSS), the satellite clock bias (SCB) plays an important role in the application of real-time precise point positioning (RT-PPP). Based on the operation of Beidou satellite global service, it is very important to establish a reliable Beidou SCB prediction model. In this research, an attention mechanism-based long short-term memory neural network (AttLSTM) model is applied to SCB prediction. The attention mechanism introduced in modelling can make the model pay less attention to useless information through weight allocation. In this paper, the BeiDou-3 Navigation Satellite System (BDS-3) satellite precision clock product provided by GFZ is used for clock prediction experiments. The proposed AttLSTM model, long short-term memory neural network (LSTM) model and quadratic polynomial (QP) model are compared and evaluated, and 12h and 24h SCB prediction experiments of BDS-3 satellite are set up. The results show that AttLSTM model can achieve high SCB prediction accuracy, and the averaged prediction accuracy of 12h and 24h can reach 1.41ns and 1.75ns. Compared with LSTM and QP models, the prediction accuracy of AttLSTM model is improved by 26.1%, 38.4% for 12h and 29.1%, 43.1% for 24h, respectively. Then, the clock bias predicted by the three models is applied to the static PPP positioning experiment, respectively. Through the analysis of the positioning results of 15 MGEX stations, the averaged positioning accuracy of AttLSTM model in the East, North and Up directions can reach 0.074m, 0.019m and 0.154m, respectively. Compared with LSTM and QP models, the positioning accuracy of AttLSTM model is improved by 42.5% and 44.4% in the East direction, 44.7% and 58.9% in the North direction, and 21.7% and 21.8% in the Up direction.
This research aims at further completing our novel Generic Multisensor Integration Strategy (GMIS) with the systematic development of three alternate attitude models, i.e., roll-pitch-heading (RPH), direction cosine matrix (DCM), and quaternion. The GMIS' potential for a true sensor level data fusion is leveraged to its full extent here by facilitating comprehensive error analysis framework in Kalman filtering. A comparative analysis between the solutions resulted from the GMIS associated with each attitude model have been analysed and compared through real road test data. The attitude models were found to perform very consistently, exhibiting the same behaviours in the residuals of the process noise and measurement vectors along with the estimated variance components. Besides, an analysis was conducted to investigate how each attitude model reacts to a sudden trajectory variation captured by the IMU. Each attitude model still performed consistently, but the DCM model in particular exhibited resistance to absorbing erroneous observations into its process noise estimates.
Soil moisture (SM) plays a vital role in agriculture, ecosystem functioning, water conservation, weather predictions and climate models. High spatial and temporal frequency data of soil moisture is crucial for agricultural and other important applications. Recent advancements have brought attention to the possibility of using GNSS reflectometry (GNSS-R) for applications on land such as snow sensing, soil moisture retrieval, sea surface monitoring and other applications in addition to positioning, navigation, and timing applications of GNSS. Cyclone Global Navigation Satellite System (CYGNSS) is designed to improve hurricane forecasting by studying the interaction between the ocean and the atmosphere within tropical cyclones. However recent studies show the opportunity of this system for high spatio-temporal soil moisture retrieval. This study presents a machine learning-based approach to get SM at a selected region in Ethiopia using CYGNSS data and analysis of the result. Artificial Neural Network (ANN) model is developed and used to predict soil moisture. The Soil Moisture Active Passive (SMAP) global soil moisture data have been used as reference data in the ML algorithm. The proposed approach has achieved a good correlation between predicted values of soil moisture and reference values from SMAP.
Geomagnetic storms, which follow solar activities like solar flares, coronal mass ejections, and high-speed solar wind streams, are significant disturbances in the global space environment. These storms occur when high-speed plasma clouds, generated by solar activity, reach the vicinity of Earth a few days later, causing disruptions in the Earth's magnetic field. This phenomenon is known as a geomagnetic storm (Gonzalez et al., 1994). Geomagnetic storms have a profound impact on GPS Precise Point Positioning (PPP) by amplifying and varying ionospheric delays in GPS phase and code data. This, in turn, affects high-precision GPS relative positioning (Odijk, 2001). In low-to-mid latitudes, geomagnetic storms can even cause disruptions in total electron content (TEC) and result in satellite signal loss (Astafyeva et al., 2014).
Since the release of Android version 7 in 2016, the smartphone users have had access to the raw global navigation satellite system (GNSS) measurements (i.e., pseudorange, carrier-phase, Doppler, and carrier-to-noise density ratio (C/N0)) through the new application programming interface (API) called android location (API level 24). This capability opens opportunities to apply different positioning techniques, ranging from absolute to differential techniques, to the smartphone observations. Precise point positioning (PPP) is a powerful method for conducting accurate real-time positioning using a single receiver, and it can be applied to the smartphone observations as well. Most PPP smartphone positioning studies have so far focused on utilizing the GNSS only observations obtained from the smartphone's API. However, incorporating additional information as constraints, such as height information, can enhance accuracy and overall stability. Although the vertical positioning accuracy of GNSS is generally lower than the horizontal accuracy, utilizing recorded height from the smartphone GNSS chipset can still be beneficial. This incorporation increases the degree of freedom and strengthens the geometry between the receiver and satellites. In this study, we assess the effectiveness of the uncombined PPP (UPPP) model in the presence of height constraints. We utilize both pedestrian walking and vehicular datasets collected by a dual-frequency Xiaomi Mi8 device to evaluate the effect of adding height constraint to PPP model. The results demonstrate an average improvement of 22% and 26% on the root-mean-square (RMS) of horizontal error and the 50th percentile error, respectively, when employing the height constraints UPPP model. Additionally, the findings indicated a decrease in PPP convergence time, further supporting the positive impact of incorporating height constraints.
The deformed or vibratory behaviors will exceed the threshold of building under the influence of external factors, so that it is necessary to monitor the variety of deformed body. Accelerometer is widely used in deformation monitoring due to small size and high sampling rate. In this paper, the fractional Kalman filter is introduced to update the accelerometer data. The influence of the order of different fractional derivatives on the filtering results of the accelerometer is studied and compared. The results show that when the system noise and measurement noise are fixed, using different derivative orders and comparing the filtering results under different derivative orders, the root mean square error of the fractional filtering model is smaller. Compare the filtering results under different noise variances. By comparing the errors of the two models, the image shows that the fractional Kalman filter model has better filtering performance than the standard Kalman filter model.
Improving the accuracy of Terrestrial Mobile LiDAR (TML) data has been a challenge in Engineering Surveys. This research aims at how to innovatively enhance the accuracy of TML solutions through post-processing toward meeting high accuracy specifications in Engineering Surveys. Three techniques are described and implemented. Firstly, the linear feature-enhanced 3D Conformal Coordinate Transformation (3DCCT) is developed by employing ground control points (GCPs) together with linear feature constraints. Secondly, a two-stage Multistrip Adjustment (MA) technique is proposed that first co-register the overlapped TML strips using tie points and tie features extracted from them and then adjust the co-registered LiDAR data by applying the feature enhanced 3DCCT. Lastly, a post-processing technique for calibrating the LiDAR boresight errors of a terrestrial LiDAR system is tested out by using its own point clouds. Their usage has been strategically studied through their applications to field-test data. Specifically, multiple scenarios have been tested, analysed, and compared in terms of the usage of GCPs, the effect of feature constraints, MA and the effect of boresight error compensation etc. As shown from the results, their utilization is encouragingly contributing to the accuracy improvement of TML data towards the high accuracy demand for Engineering Surveys. A practical implementation dataflow is outlined at the end of this manuscript.
lonospheric information plays a signifcant role in modern communication and navigation systems, This article provides a comprehensive survey of the application, modelingmethods,and results related to ionospheric infor-mation, The article frst introduces prerequisiteknowledge of the ionosphere,and then describesthe methods and techniques used in the extrac-tion of ionospheric information, the generationof ionospheric Vertical Total Electron Content(VTEC) maps, the modeling and interpolation ofthe ionosphere, and the forecasting of ionosphericinformation, The article also provides illustrativeexamples and fgures to demonstrate the effectiveness of the presented methods. Our surveyprovides insights and guidance for researchers andpractitioners interested in developing ionosphericmodeling and forecasting methods for GNSS applications.
