Leveraging carrier phase observations within Global Navigation Satellite Systems receivers allows centimeter-level positioning accuracy. However, carrier phase observations are significantly affected by additive noise, which is assumed to follow a von Mises distribution, thereby degrading the performance of phase-based positioning estimators. To improve the modeling of carrier phase observations, we propose a novel approach that constrains the parameters of the von Mises distribution-specifically, the angular location modeling the phase and its dispersion parameter kappa modeling the noise-to evolve within the Lie group space SO(2) x R+. To estimate these parameters, we employ a Lie group maximum likelihood estimator, solved through a Newton algorithm on Lie groups. This approach demonstrates advantages in terms of robustness and precision, especially when dealing with a small number of observations, compared to traditional Euclidean-based methods.
Low Earth Orbit (LEO) mega-constellations are revolutionizing global communication by delivering high-speed internet and supporting a wide range of services. Beyond communication, recent studies have shown that LEO satellites can also provide Positioning, Navigation, and Timing (PNT) services. Leveraging existing LEO Non-terrestrial-network (NTN) for dual-purpose applications has garnered significant attention as a future-oriented approach. This paper investigates the feasibility of re-purposing a well-established satellite communication protocol, Digital Video Broadcasting Second Generation Satellite Extensions (DVB-S2X), for positioning applications within LEONTN. This paper analyses the statistical acquisition performance similar to the Global Navigation Satellite System (GNSS) to evaluate the potential of DVB-S2X for positioning estimation. This analysis can provide a baseline for trade-offs for communication protocol selection when developing a sustainable integrated navigation and communication (Nav-Comm) system within LEONTN.
Global Navigation Satellite Systems (GNSS) are the primary technology for delivering reliable positioning and timing solutions, supporting a wide range of applications across different sectors. However, their performance can be significantly degraded by interference, which gives rise to the need for robust countermeasures that ensure both accuracy and resilience. This paper presents a testbed based on the Skydel GNSS Software Simulator for conducting multi-antenna GNSS experiments under realworld conditions, enabling interference mitigation studies without the need for additional hardware. The output is adjusted and sent to the FGI-GSRx GNSS software receiver, which evaluates the impact of the beamforming algorithms across all receiver stages, providing a comprehensive assessment of their effect on positioning performance. Two well-known techniques, Power Inversion (PI) and Capon (CAP), are analyzed within the context of small receivers, with results demonstrating their potential for interference mitigation even with limited antenna arrays, thus proving their suitability for handheld devices.
Indoor Positioning has grown considerably in recent years due to technological advancements and its usefulness for product tracking, pedestrian navigation, or autonomous driving. Indoor Positioning systems comprise three main components: the supporting technology, the underlying measurements, and the position estimation algorithm/method. One measurement that has been attracting attention recently is Direction of Arrival (DoA). DoA represents the direction from which a signal is impinging on a receiver. However, DoA measurements are affected by noise, interference, multipath propagation, and obstructions in the Line of Sight (LoS). In this work, we propose an approach to mitigate these problems and improve DoA estimation by applying filtering to the phase-differences obtained from each pair of elements in Linear Antenna Arrays. The results show that the accuracy in DoA estimation using filtering benefits the most inside the Field-of-View (FoV) of -40 degrees to 40 degrees. Among the several discussed methods, Z-Score provides the best average error of 5.79 degrees inside the -40 degrees to 40 degrees.
In the positioning of the Global Navigation Satellite System (GNSS) in the urban environment, the insufficient number of direct paths from available satellites due to problems such as shelter of buildings, leading to the reduction of the accuracy of the positioning of traditional positioning methods. In typical urban scenarios, specular multipaths occupy the majority, which carry abundant environmental information. Therefore, a deep neural network based GNSS multipath-assisted positioning algorithm (DL-MPAL) is proposed in this paper. In this method, multipath signal features are transformed into learnable localization features to assist localization, to improve localization accuracy and robustness. Moreover, the algorithm innovatively adopts differential pseudorange and normalized satellite position as network inputs, and receiver position in East-North-Up (ENU) coordinate system as output, realizing end-to-end mapping from signal features to receiver position. The simulation results show that the DL-MPAL algorithm reaches an average positioning error of 0.87 m and 5.21 m in two different urbanization scenarios. Also, it has high robustness under different carrier-tonoise ratio and training set label error.
This paper proposes a method to calculate and report message repetitions in multiple slots, based on combinatorics. The method is then applied to the Galileo emergency warning message (EWM) and its authentication, given the low-bandwidth, noisy channel features of GNSS. The advantages of reporting message repetition are evaluated in terms of probability of message reception and authentication for different page error rates. The proposed method can be used for other messages, longer sequences, and other communication systems and channels with similar constraints.
This paper explores the impact of electromagnetic interference (EMI) from consumer electronics and common household appliance on the performance of GNSS receivers in smartphones. As GNSS signals are weak and susceptible to external noise, consumer appliances can introduce spurious emissions near GNSS frequency bands, leading to signal degradation. To investigate this, we conducted a series of controlled experiments simulating typical smartphone usage in proximity to different sources of EMI among household appliances and consumer electronics (a microwave oven, a Wi-Fi router, and Bluetooth accessories such as gimbal and earphones). Our analysis reveals that the microwave oven can induce severe GNSS signal degradation, with substantial AGC and C/N ${ }_{0}$ drops and significant satellite availability loss, particularly in close-range configurations. Additionally, Wi-Fi interference resulted in moderate reductions in signal quality and limited-but non-negligible-inaccuracies in position estimates. In contrast, Bluetooth exhibited negligible effects under tested conditions. These findings suggest potential implications for indoor navigation applications and underscore the importance of further improvements in interference mitigation for consumer GNSS devices.
Indoor localization using fingerprinting has gained attention due to its practicality, but conventional fingerprinting methods require extensively labeled data, making large-scale deployment challenging. Few-shot learning (FSL) enables generalization from limited samples, but its application to indoor localization remains difficult due to data scarcity, distribution shifts, and privacy concerns. This paper introduces Federated Few-Shot Zone Learning (FFZL), the first framework to integrate federated FSL into fingerprinting-based indoor localization. FFZL leverages prototypical networks within a federated setting, allowing decentralized devices to collaboratively classify unseen zones with minimal labeled data while preserving privacy. To further enhance localization accuracy, we introduce FFZL-Map, an improved variant that incorporates spatial regularization into the loss function with the help of the floor map, ensuring learned representations align with the environment's geometric structure. We evaluate FFZL and FFZL-Map on both simulated and real-world datasets, demonstrating the effectiveness of our approach in adapting to novel zones with minimal supervision.
Many industrial and scientific applications rely on high accuracy positioning. With the advancements of modern Global Navigation Satellite Systems (GNSS) constellations and the release of new services, access to a real-time high-accuracy positioning anywhere in the world is a reality. To enable real-time Precise Point Positioning (PPP), Galileo High Accuracy Service (HAS) offers corrections without needing any additional infrastructures on the user side. In this paper, we compare the user positioning performance obtained with HAS corrections received via Internet Data Dissemination (IDD) and Signal-In-Space (SIS) using the software JRC User Navigation Engine (JUNE), which implement the Galileo HAS reference user algorithm developed at the Joint Research Centre (JRC). The comparison is carried out using stations of the International GNSS Service (IGS) located throughout the world as well as a receiver located at the JRC site in Ispra, Italy. The goal of the paper is to evaluate the achievable performance using the different HAS corrections and and review the future developments of the JUNE software.
GNSS is prone to jamming and spoofing, which should therefore be handled to avoid PNT errors. In this work, we propose a GNSS spoofing cancelation algorithm in a loosely INS coupled navigation filter. The proposed algorithm is based on detection of the spoofing attack, using filter innovation within a time interval, followed by determination of a point in time to recompute the navigation filter to avoid spoofing data inclusion and filter uncertainty due to rapid movements. With these algorithm attributes the same decision for detecting spoofing can be reused for deciding when a spoofing attack has ended, since more realistic covariance estimates and smaller error should be achieved even for late detected spoofing attacks. We show by simulation that the proposed algorithm perform well in relevant spoofing scenarios. It is shown that recomputation of the filter from an earlier timestep is crucial to be able to determine when the attack ends. By re-computation with INS only, and excluding GNSS when the motion is rapidly changing and when spoofing is present, slower growing errors and more realistic covariance estimates are achieved.
The prospect of using Low Earth Orbit - Position-Navigation-Timing (LEO-PNT) for indoor applications has been gaining traction in the past years, given the difficulties that Global Navigation Satellite Systems (GNSS) face in these scenarios. Inherently, LEO satellites deliver higher received power than GNSS, assuming similar carrier frequency bands, which translates to better indoor penetration capabilities and the potential of achieving an indoor PNT solution. This paper presents a comparison between various LEO-PNT system designs, comprising the choice of constellation, carrier frequency, and transmission power, and compares their performance to traditional GNSS constellations. Concise system design takeaways are provided given these extensive simulations which shed light on the requirements for LEO systems to provide the PNT solution indoors. These results show the promising potential for LEO-PNT indoors, particularly for carrier frequencies below 1.5 GHz and using the medium-sized 'Celikbilek' constellation, along with the challenges to work at higher frequencies such as 4 GHz or 7 GHz, which are envisaged for outdoor applications.
This work presents the probability of a Galileo receiver calculating a position fix and an OSNMA (Open Service Navigation Message Authentication) position fix in different environments, including discontinuous signal reception. To complement real-data tests already available, the results are based on simulations using AWGN (Additive White Gaussian Noise) and simplified LMS (Land Mobile Satellite) channels, combined with assumptions on satellite visibility to model opensky, urban and hard urban environments. TTFF (Time To First Fix) and TTFAF (Time To First Authenticated Fix) are compared for each environment for a 10 -minute window, corresponding to the usual navigation refresh rate. While the results focus on corner cases where the TTFF is achieved within the window, but not TTFAF, identifying the reasons, in most cases TTFF and TTFAF are both achieved.
Radio-based simultaneous localization and mapping (SLAM) has the potential to provide precise localization and environmental sensing capabilities using millimeter wave (mmWave) signals. In this paper, we propose methods that address the robustness and computational complexity issues of existing bistatic snapshot radio SLAM algorithms. We introduce multi-hypothesis Bayesian approaches to enhance the robustness and accuracy of solving the SLAM problem. In addition, we introduce effective methods to reduce computational complexity using prior information. The developed methods are evaluated using experimental mmWave data using 5G waveforms and benchmarked with respect to state-of-the-art methods. The results imply that the proposed methods improve the accuracy, robustness, and efficiency of radio SLAM.
Multi-modal transport refers to multiple transportation means (e.g., car, plane) that can be used to transport people or goods. Classifying the mode of transportation can have multiple usages towards sustainable transport solutions, such as optimizing routes, reducing transit times, having efficient logistics operations, reducing transportation costs by strategically combining different modes, or understanding how people move within cities for migration studies. Multi-modal transport classification has traditionally relied on data collected from various movement sensors (e.g., accelerometers, pedometers, gyroscopes); yet, with the opening of the access to raw Global Navigation Satellite Systems (GNSS) data on mobile devices, new avenues of multi-modal analysis have been created, when GNSS signals alone (without additional sensors) could be used to classify the mode of transport. This paper introduces a novel pseudorangebased approach for multi-modal transport classification, where only the instantaneous raw navigation data from two strong satellites in view is used to classify the user transportation mode at that instant. We validate our concept based on Machine Learning (ML) algorithms with data collected with four Android devices (three smartphones and a smartwatch) in 24 scenarios, encompassing five different transportation modes.
Smartwatches currently lack a precise solution for determining a user's position in emergency scenarios, limiting their effectiveness in critical situations. To address this, this study evaluates Assisted-GPS (A-GPS) with the LTE Positioning Protocol (LPP) to enhance emergency call positioning in Voice over LTE (VoLTE) systems. By leveraging network assistance, AGPS improves accuracy and reduces Time to First Fix (TTFF) compared to traditional GPS. We develop custom source code and use laboratory testbed to simulate real-world emergency call scenario, assessing key metrics such as horizontal positioning accuracy of A-GPS compared to Cell-ID, jitter, message exchange consistency via LPP, signal-to-noise ratio (SNR), network response times, and vertical positioning accuracy. The results show that A-GPS using the LPP protocol improved the location accuracy, achieving an average positioning error of 1.65 meters, in contrast to approximately 700 meters associated with Cell-ID and other methods from the literature. Additionally, the smartwatch response times were approximately 18 seconds meeting the 3GPP 20-second standard for location reporting. These findings highlight the potential of A-GPS and LPP protocol to enhance emergency call positioning.
Global Navigation Satellite Systems (GNSS) are an integral part of modern technology, providing essential positioning, navigation, and timing (PNT) services for various applications. These systems face escalating threats from sophisticated radio frequency (RF) interference, particularly linear frequency-modulated (LFM) chirp signals, which exploit rapid time-frequency variations and high bandwidth occupation to disrupt positioning, navigation, and timing services. These interference signals pose critical risks by overwhelming GNSS receivers and degrading their position accuracy. Effective mitigation techniques are therefore sought in order to protect the reliability of GNSS receivers and applications. Novel mitigation methods such as Fractional Fourier Transform (FrFT)-based techniques emerge as potential candidates to suppress chirptype interferences, but face significant computational challenges for their practical implementation. This paper addresses this limitation by introducing a structured analytical approach that establishes a direct relationship between the characteristics of the chirp signal, specifically the chirp rates, and the optimal transformation parameters in the Fourier fractional domain. The proposed implementation is shown to achieve a robust interference mitigation with minimal computational overhead, making it suitable for most GNSS applications. The findings highlight the potential for improving the resilience of GNSS receivers under RF interference and suggest avenues for future research to extend this approach to more complex signal environments.
Providing a reliable position solution is essential for many safety-critical and reliability-critical location-based services. However, the Global Navigation Satellite System (GNSS) has inherent vulnerability and satellite signals are prone to be disturbed by surrounding obstacles, leading to misleading solutions. The Solution Separation (SS) detector can cope with various fault conditions but its associated subset number is large, making the implementation process complex. To address this, a simplified solution separation method is proposed in this paper. To reduce the computational complexity, the relationship between the full set solution and the subset solution is derived in detail to avoid recalculating the subset solution from scratch. Meanwhile, a new multiple-fault exclusion scheme with fewer subsets is designed based on the above relation formula. A simulated experimental platform is built, where single and double GNSS faults are injected into the raw observations with various amplitudes. The test statistic variation is analyzed and all injected faults are excluded successfully. The position error comparison further validates the effectiveness of the proposed method.
The growing reliance on global navigation satellite systems (GNSS) for critical applications makes them a prime target for spoofing attacks. We propose a machine learning (ML)based blind authentication framework that integrates GNSS signals with low Earth orbit (LEO) signals of opportunity (SOOP), whose positions are not known a priori. Our approach leverages the consistency between GNSS and LEO-derived measurements, with a focus on Doppler shift, to detect spoofing attacks. We introduce a novel SOOP transmitter identification process based on pseudorange and Doppler shift residuals and evaluate multiple ML-based detectors, including support vector machines (SVM), one class SVM (OCSVM), and neural network (NN). Simulation results demonstrate that the proposed framework significantly improves spoofing detection performance, even with limited trusted measurements.
Indoor positioning based on 5G data has achieved high accuracy through the adoption of recent machine learning (ML) techniques. However, the performance of learning-based methods degrades significantly when environmental conditions change, thereby hindering their applicability to new scenarios. Acquiring new training data for each environmental change and fine-tuning ML models is both time-consuming and resource-intensive. This paper introduces a domain incremental learning (DIL) approach for dynamic 5G indoor localization, called 5G-DIL, enabling rapid adaptation to environmental changes. We present a novel similarity-aware sampling technique based on the Chebyshev distance, designed to efficiently select specific exemplars from the previous environment while training only on the modified regions of the new environment. This avoids the need to train on the entire region, significantly reducing the time and resources required for adaptation without compromising localization accuracy. This approach requires as few as 50 exemplars from adaptation domains, significantly reducing training time while maintaining high positioning accuracy in previous environments. Comparative evaluations against state-of-the-art DIL techniques on a challenging real-world indoor dataset demonstrate the effectiveness of the proposed sample selection method. Our approach is adaptable to real-world non-line-of-sight propagation scenarios and achieves an MAE positioning error of 0.261 meters, even under dynamic environmental conditions. Code: https://gitlab.cc-asp.fraunhofer.de/5g-pos/5g-dil
Distributed learning and Edge AI necessitate efficient data processing, low-latency communication, decentralized model training, and stringent data privacy to facilitate real-time intelligence on edge devices while reducing dependency on centralized infrastructure and ensuring high model performance. In the context of global navigation satellite system (GNSS) applications, the primary objective is to accurately monitor and classify interferences that degrade system performance in distributed environments, thereby enhancing situational awareness. To achieve this, machine learning (ML) models can be deployed on low-resource devices, ensuring minimal communication latency and preserving data privacy. The key challenge is to compress ML models while maintaining high classification accuracy. In this paper, we propose variational autoencoders (VAEs) for disentanglement to extract essential latent features that enable accurate classification of interferences. We demonstrate that the disentanglement approach can be leveraged for both data compression and data augmentation by interpolating the lower-dimensional latent representations of signal power. To validate our approach, we evaluate three VAE variants - vanilla, factorized, and conditional generative - on four distinct datasets, including two collected in controlled indoor environments and two real-world highway datasets. Additionally, we conduct extensive hyperparameter searches to optimize performance. Our proposed VAE achieves a data compression rate ranging from 512 to 8,192 and achieves an accuracy up to 99.92
