This study investigates the effect of Pedestrian Dead Reckoning (PDR) drift and radio map resolution on the performance of Surface Correlation (SC)-based indoor positioning. SC utilizes spatially accumulated Received Signal Strength Indicator (RSSI) sequences to improve localization accuracy, but its performance is sensitive to PDR errors and radio map granularity. To mitigate the impact of drift-induced misalignment, a ±10° rotation-based correction method is proposed. Simulations are conducted under varying resolutions (1 m, 2 m, 3 m) and drift levels (0.1°, 0.3°, 0.5°) to evaluate positioning accuracy and computational efficiency. The proposed method demonstrates consistent performance improvement under all drift conditions and is especially effective with high-resolution radio maps. The findings provide practical guidance for designing SC-based positioning systems with optimized trade-offs between accuracy and efficiency.

Reliable integer ambiguity resolution is essential for high-precision, high-integrity Real Time Kinematic (RTK) positioning. The probability of correct fix (P(CF)) provides a key metric for assessing ambiguity resolution reliability and is directly influenced by the integer decorrelation matrix (Z-matrix) performance. This paper proposes an enhanced Z-matrix search method to improve the reliability of integer ambiguity estimation. The method extends the conventional LAMBDA algorithm by generating multiple Z-matrix candidates through systematic linear transformations and expands the search space to select the most reliable candidate based on P(CF). Simulation results from 24-hour single-epoch RTK show the proposed method improves P(CF) in over 97.5% of epochs, reduces the average 1-P(CF) value by approximately 64.1%, and increases fixed solution availability from 85.7% to 95.3%. Although the validation was performed using single-epoch RTK, the proposed method applies to general RTK architectures, including Kalman filter-based systems, and should contribute to faster convergence and more robust ambiguity resolution.

Centimeter-level precise positioning based on satellite navigation requires satellite clock offset accuracy at the nanosecond level, which is more precise than the time information provided by broadcast ephemeris. The development of precise real-time clock error estimation technology is therefore essential. This study developed a real-time Global Positioning System (GPS) satellite clock offset estimation algorithm based on a sequential filter. The algorithm estimates satellite clock offsets using data acquired from a global network of stations. The algorithm estimates the satellite clock offset, drift, receiver clock error, tropospheric zenith delay correction, and ambiguity parameters as state variables. The observation model uses un-differenced observations with an ionosphere-free combination technique. To improve the numerical stability of the Kalman filter, Cholesky decomposition and the Joseph stabilized form of the covariance update method were introduced. For validation, the results were compared with the final clock offset product from the Jet Propulsion Laboratory (JPL). The root mean square error (RMSE) was found to be 1.47 ns when using only carrier phase and 1.56 ns when using a combination of carrier phase and code. Additionally, a bias was observed depending on the GPS satellite block, with a magnitude reaching up to 2.5 ns.

Precise time synchronization is crucial for modern infrastructure but objectively evaluating the intrinsic performance of technologies such as Global Positioning System (GPS) Precise Point Positioning (PPP), Integer Precise Point Positioning (IPPP), P3, and Two-Way Satellite Time and Frequency Transfer (TWSTFT) is challenging due to the influence of local time source stability. This paper proposes a Double Difference-based Three-Cornered Hat (TCH) method to overcome this limitation. By applying this technique to analyze TWSTFT and GPS links between Korea Research Institute of Standards and Science (KRISS) and Physikalisch-Technische Bundesanstalt (PTB), we confirmed a stability of mid 10^-17 for averaging times over 2×10⁶ seconds, consistent with similar research by the National Institute of Standards and Technology (NIST). Our findings provide an important standard for evaluating high-precision links and offers a practical foundation for technology selection and system design in fields such as national time standards and defense.

In aerial vision-based navigation systems, accurately estimating the misalignment between the camera and the inertial measurement unit (IMU) is critical. However, uncertainty in the camera's intrinsic parameters can propagate to misalignment estimation and result in geometric errors. This can lead to significant degradation of vision-based navigation performance during high-altitude operations. To address this, we propose a reversal calibration method that leverages extrinsic parameters information to determine the effective range of intrinsic parameters. Based on the results of extrinsic calibration, we define an effective correction region for intrinsic parameters with respect to the physical centers of both sensors. This relationship is then integrated into a constrained optimization framework to refine the intrinsic parameters within physically valid bounds. Experiments were conducted using a camera with a focal length of 35 mm. The results demonstrate that incorporating physical constraints into the optimization process leads to more stable parameter correction, and reduced reprojection errors by up to 60.73%.

The Moon is increasingly recognized as a strategic waypoint for future Mars and deep space missions and this is driving demand for reliable Position, Navigation, and Timing (PNT) services. Traditional reliance on the Deep Space Network (DSN) provides high post-processed accuracy but limited real-time performance and may not scale with rising mission numbers. To address this, space agencies are exploring Global Navigation Satellite System (GNSS)-based and dedicated lunar navigation systems. Recent demonstrations, such as the Lunar GNSS Receiver Experiment (LuGRE) mission, have validated GNSS positioning at the Moon with kilometer-level accuracy. This study extends these efforts by simulating lunar orbit determination using multi-GNSS constellations. An Extended Kalman Filter (EKF) approach, integrating GNSS observations with orbital dynamics, is employed. The simulation results show significantly improved position accuracy and underscore EKF's potential for future lunar navigation systems.

In this study, we propose a Graph Neural Network (GNN)-based localization approach utilizing beam Reference Signal Received Power (RSRP) measurements in 5G networks. Existing GNN-based localization approaches have mainly targeted indoor environments with dense Base Station (BS) deployments and are less effective in outdoor scenarios where the number of BSs is limited. To overcome this limitation, we construct an expanded graph extracted by features such as RSRP, angle, and Line of Sight/None-Line of Sight (LoS/NLoS) indicators, assuming that such information is available at the User Equipment (UE). We validate the proposed approach through ray-tracing simulations in a realistic outdoor 5G environment and evaluate its localization accuracy compared to the triangulation approach.

Global navigation satellite system (GNSS) provides position, velocity, and timing information globally, but single-frequency receivers are limited by errors such as satellite orbit, clock, and especially ionospheric delay. While satellite-based augmentation system (SBAS) improves accuracy by supplying orbit, clock, and ionospheric corrections, realistic simulation tools are required to effectively develop and validate such systems. This study presents a MATLAB-based simulator that generates GNSS observations, computes corrections via precise orbit determination using extended Kalman filtering (EKF), and constructs grid-based ionospheric maps from dual-frequency measurements. The simulator, focused on positioning accuracy rather than integrity, is evaluated using a virtual quasi-zenith satellite system (QZSS) constellation with global positioning system (GPS) under intense ionospheric conditions in Japan. The performance of the simulator was analyzed at the MIZU ground station and further assessed across 36 virtual stations to demonstrate the spatial effectiveness of the correction information.

In various aerospace and military applications, precise location estimation techniques are essential to accurately determine the location of objects at specific points in real time. In particular, when estimating the location of a Fast-Moving Object (FMO) with rapid location changes in a short period of time, such as a high-speed aircraft or a train, an algorithm that satisfies both high accuracy and fast processing speed is required. The representative location estimation techniques include Time Difference of Arrival (TDOA) and Frequency Difference of Arrival (FDOA). TDOA provides high accuracy but it does not consider the object's travel distance during the processing time, which causes serious location errors in the FMO environment. Although FDOA efficiently estimates both location and velocity, it has the disadvantage of being extremely sensitive to the initial location estimate. In this paper, we propose a combined location estimation algorithm with TDOA and FDOA in three-dimensional space that efficiently estimates the location of FMOs and compensates the above limitations of both techniques. The proposed algorithm first estimates the initial location of a FMO employing TDOA and then calculates its velocity employing FDOA to predict the FMO's trajectory. The location estimation performance of the proposed algorithm is evaluated through computer simulations under various scenarios, including different sensors, FMO placement, speed, etc.

Global navigation satellite system (GNSS) signal delay caused by the troposphere is a significant source of error for precision positioning. To mitigate this tropospheric delay effect, a precise point positioning (PPP) method estimates the tropospheric delay with carrier phase ambiguities. With accurately predicted tropospheric delay, this delay effect can be effectively minimized or the tropospheric delay estimation process can be improved. An artificial intelligence technique, Long Short-Term Memory (LSTM), was developed to predict the wet tropospheric delay using meteorological observations such as temperature, pressure, and relative humidity at GNSS monitoring stations. Five years of meteorological observation data from four GNSS measuring stations in the Korea were used to train the LSTM network model, which predicts one year of wet tropospheric delays at those stations. The predicted wet delays are compared with the wet delays estimated by the PPP method, and the forecast accuracy and geographical error differences are discussed.

Time Difference of Arrival (TDOA) localization using GNSS-synchronized surface buoys offers a promising solution for underwater emitter tracking without direct cable or acoustic links. However, the achievable localization accuracy depends on the interplay of buoy geometry, time-synchronization errors, and dynamic tracking constraints. This paper presents a quantitative analysis of these trade-offs using both static and dynamic scenarios. For static localization, Geometric Dilution of Precision (GDOP) and Root Mean Square Error (RMSE) are evaluated under varying buoy formations and clock errors. For dynamic tracking, an Extended Kalman Filter (EKF) framework incorporating realistic ping update rates is applied to evaluate track persistence and accuracy over time. Simulation results show that the buoy geometry strongly influences GDOP patterns and error sensitivity, time-synchronization errors above the millisecond scale rapidly degrade localization performance, and in dynamic tracking, slower ping rates combined with unfavorable geometries can cause error plateaus even when filtering is applied. These findings provide practical guidelines for designing GNSS-synchronized buoy networks by balancing geometry selection, synchronization precision, and update rate according to operational requirements.

This paper proposes an efficient Regional Ionospheric Map (RIM) / Differential Code Bias (DCB) simulator based on Receiver Independent Exchange (RINEX) and IONosphere Map Exchange (IONEX) format files. Using the proposed simulator, we evaluated if it is possible to discriminate and estimate RIM and DCB variables based on a single regional satellite constellation and reference stations distributed in a restricted area. The proposed simulator embeds a recursive estimator that can discriminate and estimate RIM and DCB variables based on basic RINEX format navigation and observation files. It facilitates the evaluation of accuracy and precision by referencing the IONEX files generated by the estimator and International GNSS Service (IGS). Experiments were performed to evaluate the simulator under difficult environment by applying the RINEX data of the 4 regional satellites received by 11 reference stations distributed in Asia and Oceania region. By the experiment result, the average and the standard deviation of the vertical ionospheric delay (VID) difference were obtained as 0.35 m and 2.73 m, respectively. In the case of satellite DCB, the estimated values matched the group delay values in the broadcast ephemeris parameters with the standard deviation within 0.6 m if the unknown mean value were ignored.

Precise time synchronization among distributed infrastructure is essential for enhancing the performance of systems such as wireless sensor networks, cellular base stations, and terrestrial navigation systems. Global Navigation Satellite System (GNSS) time transfer is a widely used method for synchronizing remote receivers; however, due to communication or operational constraints, it is often performed only intermittently at periodic intervals. During these intervals, clock offsets between receivers may accumulate due to clock drift and can lead to degraded time synchronization accuracy. To address this issue, this study evaluates and compares three clock drift modeling methods-moving average, polynomial fitting, and Kalman filtering-for predicting and compensating clock offsets between GNSS receivers in the absence of continuous GNSS time transfer. Both 2-state and 3-state Kalman filter models are considered to examine how the number of state variables affects prediction performance. Experiments were conducted using real data collected from two GNSS receivers: one disciplined by an external rubidium atomic clock and the other steered to GPST. The results show that when the GNSS time transfer interval is less than 20 hours, the 2-state Kalman filter achieves the best performance, whereas the 3-state Kalman filter performs better for longer intervals. These findings provide practical guidelines for selecting appropriate clock drift compensation strategies and GNSS synchronization intervals based on timing accuracy requirements and system constraints.

This study evaluates the performance of Global Navigation Satellite System (GNSS) data processing using the Qinertia software (SBG Systems). Three processing modes-Single Base Station (SBS) Post-Processing Kinematic (PPK), Virtual Base Station (VBS) PPK, and Precise Point Positioning (PPP)-were applied to both static and kinematic positioning scenarios. The results were compared against coordinates derived from the high-precision GipsyX scientific software as well as real-time high-accuracy methods including Real-Time Kinematic (RTK) and Virtual Reference Station (VRS) RTK. For static positioning, all three methods achieved high accuracy, with average horizontal and vertical root mean square error (RMSE) of 0.8 cm and 1.5 cm, respectively, relative to GipsyX and RTK reference solutions. Notably, VBS PPK maintained a fix rate of approximately 99% for baseline distances exceeding 30 km. In kinematic positioning, both SBS-PPK and VBS-PPK achieved centimeter-level accuracy compared to VRS-RTK. Under open-sky conditions, average horizontal and vertical RMSE were 3.5 cm and 2.5 cm, respectively, while in dense urban environments they were 3.0 cm and 5.8 cm. Performance differences between the two PPK methods under open-sky conditions were minimal, within about 1-2 cm.