Personal safety and crime prevention have become pressing societal concerns. While wearable devices such as smartwatches offer features including Global Positioning System (GPS) tracking and emergency alerts, their ability to proactively recognize deviations from a user's usual path is limited. This study proposes a Long Short-Term Memory (LSTM) based trajectory learning algorithm that leverages anonymized user data without additional identifiers. It enables detection of changes from a user DB of usual trajectories and thus allows recognition of anomalies in real-time. Experimental results demonstrate that the model achieves relatively consistent performance in predicting distance errors for paths, although the time prediction performance may vary depending on path characteristics. In anomaly detection analyses, normal paths maintained stable values without exceeding the set threshold, while anomalous paths exhibited increasing error values over time, eventually exceeding the threshold.

With the availability of smartphone GNSS raw measurement data, research on enhancing positioning accuracy using smartphones has become increasingly prominent. Despite the widespread use of C/N₀-based weighting models to mitigate multipath errors, these models do not sufficiently account for the intrinsic characteristics of the observations, resulting in unstable positioning accuracy. To address this limitation, this study proposes a novel weighting model that incorporates the inherent characteristics of GNSS observations by combining code pseudorange residuals and C/N₀. The proposed model utilizes a sliding window technique to calculate residuals through curve fitting of code pseudorange observations. These residuals are then combined with C/N₀ to better reflect the unique features of the observations. Experimental results from tests conducted at Inha University demonstrate significant improvements in positioning accuracy. The proposed model achieved 40%, 32%, and 33% improvements in horizontal positioning accuracy and 32%, 20%, and 23% improvements in vertical accuracy for DOY 022, DOY 058, and DOY 253, respectively, compared to conventional elevation-based and C/N₀-based models. These findings confirm that the proposed approach effectively improves both the accuracy and precision in smartphone GNSS positioning by addressing both observation noise and signal quality.

The accuracy of tropospheric delay corrections for the global navigation satellite system (GNSS) depends on the quality of the tropospheric model. The empirical tropospheric models used in GNSS processing include the Saastamoinen, global pressure and temperature (GPT), and global mapping function (GMF). In the present study, we estimate precise point positioning (PPP) zenith total delay (ZTD) using GPT, GPT2, and GPT3 empirical models, and then compare the results. To verify the PPP ZTD obtained from the GPT model, we compared it with the international GNSS service (IGS) ZTD products. The root mean square (RMS) value at the DAEJ station was estimated to be 4.97 mm. This is close to the accuracy of the IGS ZTD, which is about 4 mm. As a result, it is suggested that the PPP ZTD estimated in this study is reliable. In addition, we confirmed that there is a bias of 0.33 mm, 0.26 mm, and 0.49 mm between the GPT and other models at the DAEJ, MIZU, and DARW stations, respectively. On the other hand, no bias was observed between the GPT2 and GPT3 models, except for at DARW, and their ZTD values were in good agreement.

In Global Navigation Satellite System (GNSS) signal transmission, the ionosphere is the most significant source of error. The Klobuchar model, a widely used ionospheric error correction model, relies on correction coefficients transmitted in Global Positioning System (GPS) navigation messages to provide global coverage. However, its accuracy is limited to approximately 50-60%. In this study, correction coefficients and Delay Constant (DC) values were estimated using International GNSS Service (IGS) Global Ionospheric Map (GIM) data to enhance the accuracy of the regional Klobuchar model, and the resulting improvements were analyzed. The grid point range for estimating the correction coefficients was defined as latitude 22.5-50° and longitude 105-150°, considering the Ionosphere Pierce Point (IPP) range over the Korean Peninsula. After calculating the Vertical Total Electron Content (VTEC) for each model, correction coefficients and DC values were simultaneously estimated using the least squares method. To evaluate the performance, the DC values were analyzed over periods with varying solar activity, and accuracy was validated based on VTEC Root Mean Square Error (RMSE) and positioning accuracy using IGS GIM as a reference. When DC values were estimated simultaneously, the VTEC RMSE improved by 54.8-74.4% compared to the original Klobuchar model, regardless of the solar activity intensity. Furthermore, positioning results at the Daejeon Station indicated horizontal accuracy improvements of approximately 1.7-40.0% and vertical accuracy improvements of approximately 34.8-37.4%.

The ranging error of the global navigation satellite system (GNSS) caused by the ionosphere significantly degrades GNSS positioning accuracy. Correcting and predicting ionospheric delay is thus important for improving GNSS accuracy. The ionospheric delay mainly depends on solar activity. Accurate prediction of the F10.7 index, an essential indicator of the intensity of solar activity, is required to accurately predict the global ionospheric delay. Two time series forecasting algorithms have been developed for the F10.7 prediction in this study: the auto-regressive moving average (ARMA) model and the long short-term memory (LSTM) model. The predictions were performed for one-day and seven-days ahead data. The LSTM model was trained with 45 years of data and the ARMA model was trained with 270 days of data. For the one-day prediction, both the ARMA and LSTM methods yielded a low prediction error, less than 2.1% of the true F10.7 value. But for the seven-day prediction, the LSTM method yielded a lower level of prediction error, 7.1%, than the ARMA method.

The Wide-Area Differential Regional Navigation Satellite System (WAD-RNSS) is a satellite navigation system that applies a wide-area differential navigation algorithm, previously utilized in Global Navigation Satellite Systems (GNSS), to Regional Navigation Satellite Systems (RNSS). WAD-RNSS broadcasts both navigation signals and correction signals directly from regional navigation satellites and provides integrity alerts to users in the event of integrity threats by utilizing these signals. One of the most critical metrics for evaluating navigation performance, integrity refers to the system's capability to preemptively warn users of system malfunctions or significant errors in position estimation. By transmitting navigation messages, correction messages, and Non-Standard Code (NSC) signals, WAD-RNSS ensures robust integrity through the dissemination of integrity alerts. This paper analyzes the limitations of integrity alert mechanisms in conventional satellite navigation systems and introduces the integrity alert methods and operational concepts of WAD-RNSS, which combines the characteristics of Global Positioning System (GPS) and Satellite-Based Augmentation Systems (SBAS). The proposed system highlights the potential to reduce the system response and recovery time following integrity alerts compared to existing satellite navigation systems.

Satellite navigation provides positioning and timing services through Signal-In-Space (SIS) signals. The User Range Accuracy (URA) is a conservative standard deviation estimate of the SIS User Range Error (URE), which arises from errors in the navigation messages contained within the SIS signals. While the URA computation equation is well established for the GPS, it should be modified for other constellations due to differing orbital characteristics. This paper derives a URA computation equation for a Wide Area Differential Radio Navigation Satellite System covering the Korean peninsula. This system is assumed to use Inclined Geosynchronous Orbit and Geostationary Earth Orbit satellites. The derived equation accounts for the higher orbital altitudes of these satellites compared to GPS, where URA is assumed to be broadcast in the Legacy Navigation format. Simulation results confirm that the proposed URA formulation conservatively bounds SIS UREs within the service area under various user locations and prediction intervals. Additionally, URA variations for different reference station distributions are analyzed. The results show that a denser and more widely distributed reference station network leads to a smaller URA.

Global Navigation Satellite System (GNSS) operates with 20 to 40 satellites while Regional Navigation Satellite System (RNSS) is served by approximately 10 satellites. Consequently, the limited number of satellites can adversely affect stand-alone positioning. To overcome this limitation, it is important to track as many satellites as possible for as long as possible. In this study, we introduce a vector tracking loop technique that utilizes strong signals to compensate for weak signals. However, during the vector tracking loop process, there is a possibility that a weak signal may negatively affect other signals. To address this problem, an adaptive covariance matrix was applied to optimize the tracking loop structure based on the residuals between the prediction phase and the actual measurements. For the extended QZSS (consisting of three GEO and four IGSO), we applied a minimum-angle criterion to analyze the visibility in a flat environment. By generating and processing signals under conditions of varying number of visible satellites, we found that the vector tracking loop outperforms the scalar tracking loop by more than 70% in reducing the Distance Root Mean Square (DRMS) when the signal is attenuated due to the decreasing elevation angle of the IGSO satellites.

In this study, we analyze the performance of the Precise POsitioning and INTegrity monitoring (POINT) system, a Precise Point Positioning (PPP)-Real-Time Kinematic (RTK) based maritime precise Positioning, Navigation and Timing (PNT) system currently under development in Korea. The POINT system is a centimeter-level Global Positioning System (GPS) augmentation system that generates correction and integrity information based on GPS measurements received from the Continuously Operating Reference Station (CORS) of the National Maritime PNT OFFICE (NMPNT). To analyze the performance of the system, we performed static and dynamic experiments and compared the performance results with those of two existing precise positioning methods, Real-Time Kinematic (RTK) and Precise Point Positioning (PPP). From the results of the static experiment, the horizontal and vertical Root Mean Square (RMS) of POINT was 3.28 cm and 7.19 cm, respectively. From the results of the dynamic experiment in a maritime environment, the horizontal and vertical RMSs of POINT are 3.63 cm and 8.14 cm compared to RTK and 7.75 cm and 12.55 cm compared to PPP, respectively.

A service that guarantees centimeter-level positioning accuracy in domestic coastal waters is essential for advanced maritime mobility, such as autonomous ships, port logistics automation, and maritime drones. Since April 2020, the Ministry of Oceans and Fisheries has been developing the Precise POsitioning and INTegrity monitoring (POINT) service, a ground-based centimeter-level service that ensures centimeter-level positioning performance in advanced maritime mobility. The POINT system generates centimeter-level correction information by processing Global Positioning System (GPS) data collected from reference stations at a central processing center. Additionally, augmentation information is generated and broadcast using GPS data from monitoring stations. In 2023, a real-time operational system was established at the Korea Research Institute of Ships and Ocean Engineering's (KRISO) Maritime Positioning, Navigation and Timing (PNT) Operations Center. This paper presents a performance analysis conducted on the operation of the POINT service. First, performance requirement indicators and verification procedures were established, and these procedures were tested and verified by a third-party organization. During the verification process, positioning accuracy, integrity risk, availability, and continuity were evaluated across 19 monitoring stations nationwide. The positioning accuracy of the user environment was also verified on receiving platforms. The results showed that, all performance indicators were successfully met.

The performance of the Korea Augmentation Satellite System (KASS), MTSAT Satellite-Based Augmentation System (MSAS), and BeiDou Satellite Augmentation System (BDSBAS) is respectively analyzed in the Korean Peninsula and its surrounding regions. The analysis is focused on comparing ionospheric vertical delay values based on the Grid Ionospheric Vertical Error (GIVE) provided by the Satellite-Based Augmentation System (SBAS), considering regional and seasonal characteristics. The results, assuming users located in Korea, indicate that KASS provides high consistency across seasons, regions, and Ionospheric Pierce Point (IPP) inclusion rates, exhibiting excellent performance based on precise positional accuracy. MSAS shows comparable performance to KASS in low-latitude regions but displays a tendency of decreased accuracy with increasing latitude. BDSBAS provides modeling of the ionospheric vertical delay that most closely resembles the Global Ionospheric Map (GIM). However, it has the lowest IPP inclusion rate within the Korean Peninsula, indicating limited availability. Additionally, analysis of data from the DAEJ (Daejeon) and CHAN (Changchun) sites shows that ionospheric delay corrections based on GIVE provided by SBAS yield results very similar to those obtained using GIM. Notably, at the DAEJ site, SBAS outperformed GIM, and at the CHAN site, SBAS proved to be a sufficient replacement for GIM.