This article presents a tightly integrated fusion algorithm that accelerates precise point positioning (PPP) convergence and enhances localization robustness in complex urban environments. The proposed method combines a filter-based visual-inertial odometry (VIO) framework with semantic constraints from a high-definition (HD) vector map, incorporating lane-line and directional arrow features extracted via a neural network. These HD map observations provide reliable lateral and heading constraints, effectively compensating for the degraded observability and drift in traditional VIO systems. In satellite-visible conditions, they also offer accurate initial estimates that significantly accelerate PPP convergence. The system is evaluated in diverse real-world scenarios-including open roads, urban canyons, tunnels, and soundproofed elevated highways-demonstrating submeter positioning accuracy and PPP convergence within 3-10 s, representing up to a 98.33% reduction in convergence time compared to conventional PPP. Furthermore, the approach operates using low-cost inertial and visual sensors, ensuring practical deployment in large-scale autonomous navigation systems.

To address the challenges of low image contrast and distortion in low-light environments, which make it difficult to extract visual features, leading to high mismatch rates, and subsequently causing visual navigation positioning drift and degradation of the performance of global navigation satellite system (GNSS)/Vision/ inertial navigation systems (INSs) integrated systems. This paper proposes a tight-coupled global navigation and positioning method based on image enhancement. First, a multi-head self-attention image enhancement module is designed, which integrates lighting information, depth-separable convolutions, and channel attention mechanisms. A custom loss function is used to improve the quality of low-light images, achieving a peak signal-to-noise ratio (PSNR) of 23.36 dB, significantly enhancing the robustness of subsequent feature extraction and matching; Furthermore, based on the enhanced visual feature information, the GNSS, visual, and IMU raw measurements are fused to construct the FE-(feature enhanced Gnss vision inertial odometry) tight integration navigation model. Navigation experiments using EOROC and self-collected datasets demonstrate that the proposed method achieves average positioning root mean square error (RMSE) of 0.188 m, 0.192 m, and 0.246 m along the X, Y, and Z axes, respectively, with positioning residual mean values reduced by 12.3% and 25.66% compared to GVINS and VINS-Fusion, respectively, angular errors are consistently within 1.76 degrees, and the proposed method provides a technical reference for autonomous navigation of payloads in low-light scenarios.

Integrated precise orbit determination (IPOD) of Global Positioning System (GPS) satellites and Low Earth Orbit (LEO) satellites has been widely adopted to enhance the accuracy of orbit and geodetic parameters. Integer ambiguity resolution (IAR) can make full use of the inherent accuracy of carrier phase measurements, thereby serves as an essential procedure for IPOD. The expansion of operational LEO satellite number will facilitate the construction of diverse baselines for double-differenced (DD) phase ambiguity fixing, including ground-ground, LEO-LEO, and ground-LEO baselines. However, effectively combining them remains challenging due to their substantially different characteristics and contributions to IPOD. In this study, we utilize a dataset collected by a global network and seven LEO satellites, to investigate the IAR strategy for combining multiple types of baselines. Taking into account the kinematic characteristics of LEO-related baselines, we experimentally determine an appropriate minimum common-view time threshold of 10 min for LEO-related ambiguities after a comprehensive evaluation of ambiguity quantity, residuals, fixing rate, and orbital contribution. Under this strategy, the ground-LEO solution performed best in GPS orbit accuracy when using individual-type baselines, while the LEOLEO solution is least accurate. We further evaluate IAR strategies based on multi-type baselines. The allambiguity strategy outperforms the independent-ambiguity approach in the trade-off between ambiguity independence and orbital contribution. The solution combining all baseline types delivered the best performance, improving GPS orbit accuracy by 6.2% and LEO accuracy by 8.6% over the IPOD solution with ground-only IAR. The contribution of LEO-related IAR becomes more substantial in the regional scenario.

Recent developments in Multimodal Large Language Models (MLLMs) have significantly improved Vision–Language (VL) reasoning in 2D domains. However, extending these capabilities to 3D scene understanding remains a major challenge. Existing 3D Multimodal Large Language Models (3D-MLLMs) often depend on 3D data inputs, which limits scalability and generalization. To address this limitation, we propose Vid-LLM, a video-based 3D-MLLM that directly processes video inputs without requiring external 3D data, making it practical for real-world deployment. In our method, the geometric prior are directly used to improve the performance of the sceen perception. To integrate the geometric cues into the MLLM compactly, we design a Cross-Task Adapter (CTA) module to align the 3D geometric priors with the vision-language representations. To ensure geometric consistency and integrity, we introduce a Metric Depth Model that recovers real-scale geometry from the reconstruction outputs. Finally, the model is fine-tuned with a two-stage distillation optimization strategy, realizing fast convergence and stabilizes training. Extensive experiments across diverse benchmarks verified the effectiveness of our method on 3D Question Answering, 3D Dense Captioning and 3D Visual Grounding tasks, demonstrating the superior multi-task capabilities.

The dynamic approach integrates Global Positioning System and K-band range-rate (KRR) observations to enable precise orbit determination (POD) and gravity field recovery. However, background model uncertainties and temporal aliasing introduce frequency-dependent noise into the post-fit KRR residuals, thereby degrading overall solution accuracy. To mitigate these effects, empirical signals are typically modeled using either dynamic (DYN) or kinematic (KIN) parameterization strategies. Nevertheless, the combined use of DYN and KIN parameterizations remains largely unassessed, and their potential synergistic impact on POD and gravity field recovery merits systematic evaluation. This study evaluates the individual and joint impacts of DYN and KIN (DYN+KIN) on The Gravity Recovery and Climate Experiment (GRACE) Follow-On orbit accuracy and monthly gravity field recovery using nearly one year of 2019 data (excluding February due to severe data gaps). The refined solutions act as empirical temporal filters, effectively suppressing low-frequency components in KRR residuals, particularly below 1-cycle-per-revolution. Relative to nominal ambiguity-fixed reduced-dynamic orbits, the refined solutions mainly enhance the cross-track component, with DYN+KIN showing the largest improvement, while along-track precision experiences only minor (sub-millimeter) degradation. Overall three-dimensional orbit accuracy improves from 3.8 cm to 3.0 cm (DYN), 2.8 cm (KIN), and 2.8 cm (DYN+KIN). In terms of gravity field recovery, the DYN+KIN solution begins to exhibit more pronounced deviations from the other solutions beyond degree and order 30. Over oceanic regions, residual mass anomaly analysis shows that the DYN+KIN solution is associated with an approximately 16% higher noise level compared to the individual DYN and KIN strategies, which exhibit modest noise reductions relative to the nominal solution. The DYN+KIN also exhibits a dampened ~160-day periodicity in the temporal evolution of low-degree coefficients (e.g., C2,0), likely due to spectral overlap between empirical parameter frequencies and low-degree gravity signal components. These results indicate that over-parameterization introduces spectral redundancy and absorbs geophysical signals, underscoring the need to balance parameter flexibility and signal fidelity in gravity recovery strategies.

The tightly coupled integration (TCI) of GNSS Precise Point Positioning (PPP) and Micro-Electro-Mechanical Systems (MEMS) Inertial Navigation System (INS) has emerged as a promising approach for urban navigation, yet its performance degrades significantly by multipath interference common in complex urban environments. This study proposes a novel framework to enhance PPP/MEMS-INS integration through an adaptive code pseudorange classification and smoothing algorithm: (1) adaptive code pseudorange classification, where code measurements within a sliding window are statistically characterized and are used to classify subsequent data into normal/abnormal categories with adaptive weighting; (2) LSTM (Long Short-Term Memory)-based predictive modeling, which leverages initial data to train a LSTM network and smooths the abnormal code measurements of each satellite in challenging environments. Experimental results show that the proposed method reduces RMS errors in urban canyons by up to 48

Satellite attitude, affecting the non-conservative force modeling, antenna phase center and phase wind-up corrections, acts as a critical role in high precision GNSS (Global Navigation Satellite System) data analysis. Of the attitude related errors, the phase wind-up correction one is largely absorbed into the satellite clock solutions. To improve the consistency between server and user ends, the IGS (International GNSS Service) proposed to exchange satellite attitude in ORBEX (ORBit EXchange) format in 2017. However, the satellite attitude for individual satellite types vary greatly among the involved ACs (Analysis Centers). To assess the quality of satellite attitude products, most existing researches apply the method of inter-comparison between ACs. In contrast, the analysis of satellite attitude products is realized with the latest analytical satellite attitude models in this study. With the attitude products from four ACs spanning from January to June in 2024, the GPS and BDS-3 (BeiDou-3 Navigation Satellite System) MEO (Medium Earth Orbit) satellite attitude are assessed comprehensively. The results indicate: In maneuver period, aside from the block IIR satellites in GFZ AC, all of the ACs apply modeled attitude. Concerning the block IIR satellites, compared to the modeled one, the yaw differences are negligible for the CODE and GRGS ACs, however, quite large for the WHU AC. The exceeding of yaw rate than hardware limitation suggests model deficiency of WHU AC at low Sun elevation angle. As for the block IIF satellites, in comparison to the other ACs, the shadow and noon maneuvers show apparently different behaviors for the GRGS AC. With the exit time delayed or advanced, yaw jumps are observed after the noon maneuver. As for the block IIIA satellites, while the CODE/WHU ACs directly adopt the block IIR model, the GRGS AC apply the block IIF model with shadow maneuver replaced by midnight maneuver. Similar to the block IIR satellites for WHU AC, abnormal yaw rates are noticed for several block IIF/IIIA satellites at low Sun elevation angle. In the case of BDS-3 MEO satellites, the models of CAST/SECM satellites are consistent with the ESA model for the CODE/WHU ACs and similar to the ESA/CSNO models respectively for the GRGS AC. However, at low Sun elevation angle, yaw anomalies are found for the CAST/SECM satellites in GRGS AC and SECM satellites in WHU AC. While the models of GPS/BDS-3 MEO satellites can be identified for the CODE/GRGS/WHU ACs, those for the GFZ AC are generally unclear except for that of the SECM satellites is consistent with the ESA model. (c) 2025 COSPAR. Published by Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

Most of the existing indoor positioning and mapping methods based on the Implicit Neural Representations (INRS) rely on input RGB-D data with known poses, which have pose limitations and depth map holes. This paper uses stereo visual-inertial odometry (SVIO) to acquire carrier position and depth information in real-time, and integrates it with neural volume rendering (NVR) to acquire three-dimensional models with high integrity. First, a joint camera-IMU calibration method with additional geometric constraints is proposed, achieving an accuracy of better than 0.15 pixels. Second, the residual network and multi-scale cost aggregation module are combined to achieve stereo matching in low-texture environments and provide reliable depth constraints for volume rendering. Last, the joint encoding and residual feedforward multilayer perceptron are combined for volume rendering to obtain smooth and high-fidelity three-dimensional models and poses. The test results on public and self-acquired datasets show that the accuracy and completion of the reconstructed model can reach 2.25 cm and 2.21 cm on average, respectively, and the average absolute positioning errors in the X, Y and Z directions can reach 2.16 cm, 3.60 cm and 1.43 cm, respectively, which can provide reliable technical support for indoor positioning and mapping technology.

In autonomous robotic systems, precise localization is a prerequisite for safe navigation. However, in complex urban environments, GNSS positioning often suffers from signal occlusion and multipath effects, leading to unreliable absolute positioning. Traditional mapping approaches are constrained by storage requirements and computational inefficiency, limiting their applicability to resource-constrained robotic platforms. To address these challenges, we propose 3DGS-LSR: a large-scale relocalization framework leveraging 3D Gaussian Splatting (3DGS), enabling centimeter-level positioning using only a single monocular RGB image on the client side. We combine multi-sensor data to construct high-accuracy 3DGS maps in large outdoor scenes, while the robot-side localization requires just a standard camera input. Using SuperPoint and SuperGlue for feature extraction and matching, our core innovation is an iterative optimization strategy that refines localization results through step-by-step rendering, making it suitable for real-time autonomous navigation. Experimental validation on the KITTI dataset demonstrates our 3DGS-LSR achieves average positioning accuracies of 0.026m, 0.029m, and 0.081m in town roads, boulevard roads, and traffic-dense highways respectively, significantly outperforming other representative methods while requiring only monocular RGB input. This approach provides autonomous robots with reliable localization capabilities even in challenging urban environments where GNSS fails.

With the rapid advancement of autonomous driving technology, achieving accurate and continuous navigation in complex urban environments has become a critical research area. However, challenges such as dynamic objects and complex lighting conditions can significantly compromise the accuracy of visual-inertial odometry. In addition, structures such as tunnels, overpasses, and urban canyons may disrupt the continuity and accuracy of global navigation satellite system (GNSS) signals. This article presents a tightly coupled model that integrates GNSS, inertial navigation system (INS), vision observations, and lightweight map priors. This system leverages prior trajectory information from the maps to generate discrete control points, ensuring adherence to designated roads. It also utilizes deep learning algorithms to identify and exclude dynamic objects, thus enhancing the reliability of visual odometry calculations. This approach maintains decimeter-level accuracy even during prolonged GNSS signal outages and facilitates rapid convergence of precise point positioning (PPP). In urban scenarios featuring one long tunnel, our algorithm demonstrated robustness with an overall root mean square error (RMSE) of 0.9 m, showcasing its effectiveness in challenging conditions with frequent GNSS disruptions. In environments without tunnels, the accuracy is further improved, achieving an RMSE of 10 cm, which surpasses the accuracy of standalone PPP solutions.

Accurate, continuous, and reliable positioning is critical to achieving autonomous driving. However, in complex urban canyon environments, the vulnerability of stand-alone sensors and non-line-of-sight (NLOS) caused by high buildings, trees, and elevated structures seriously affect positioning results. To address these challenges, a sky-view image segmentation algorithm based on a fully convolutional network (FCN) is proposed for NLOS detection in global navigation satellite systems (GNSSs). Building upon this, a novel NLOS detection and mitigation algorithm (named S−NDM) uses a tightly coupled GNSS, inertial measurement units (IMUs), and a visual feature system called Sky−GVIO with the aim of achieving continuous and accurate positioning in urban canyon environments. Furthermore, the system combines single-point positioning (SPP) with real-time kinematic (RTK) methodologies to bolster its operational versatility and resilience. In urban canyon environments, the positioning performance of the S−NDM algorithm proposed in this paper is evaluated under different tightly coupled SPP−related and RTK−related models. The results exhibit that the Sky−GVIO system achieves meter-level accuracy under the SPP mode and sub-decimeter precision with RTK positioning, surpassing the performance of GNSS/INS/Vision frameworks devoid of S−NDM. Additionally, the sky-view image dataset, inclusive of training and evaluation subsets, has been made publicly accessible for scholarly exploration.

With the capability for single-receiver AR (Ambiguity Resolution), real-time GNSS (Global Navigation Satellite System) satellite integer clock has been the research hotspot of GNSS community in the past decade. Robust ambiguity datum is one prerequisite for high-quality integer clock. In this paper, with undifferenced AR model derived, a robust ambiguity datum definition method is proposed to isolate the signal biases from the integer ambiguity. Firstly, based on the ambiguity graph, the ambiguity datum is selected with the minimum spanning tree algorithm. Secondly, the ambiguity datum is transferred to the subsequent epochs through signal bias connection. Moreover, additional datum ambiguities are introduced to accommodate the arising of receiver and satellite in real-time circumstance. With the combination of ambiguity graph and signal bias, the efficient definition and smooth transition of ambiguity datum are assured for the real-time processing. To validate the approach, simulated real-time integer clock estimation for GPS/Galileo/BDS-3 satellites are conducted with one-month (January, 2022) data from about 100 global IGS (International GNSS Service) network stations. Experiments show that the proposed robust ambiguity datum can produce accurate real-time integer clock. The average WL (Wide-Lane) and NL (Narrow-Lane) fixed rates exceed 99% and 95%, respectively, for most of the GPS/Galileo satellites. While comparable WL fixed rate is achieved for the BDS-3 satellites, the NL fixed rate is about 5% lower as a result of poor orbit quality and imperfect measurement model. With eclipsing satellites excluded, the standard deviations between real-time integer clock and post-processed reference clock reach 0.018 ns and 0.016 ns for the GPS and Galileo constellations, respectively, which are 28% and 24% smaller than those of corresponding float clocks. In contrast, the standard deviations of float and integer clocks are similar for the BDS-3 constellation. Further improvements on the background models, including unconservative force modeling, satellite attitude modeling and antenna calibration are required to achieve ultimate accuracy, especially for the BDS-3 satellites.

Precise, consistent, and reliable positioning is crucial for a multitude of uses. In order to achieve high precision global positioning services, multi-sensor fusion techniques, such as the Global Navigation Satellite System (GNSS)/Inertial Navigation System (INS)/Vision integration system, combine the strengths of various sensors. This technique is essential for localization in complex environments and has been widely used in the mass market. However, frequent signal deterioration and blocking in urban environments exacerbates the degradation of GNSS positioning and negatively impacts the performance of the multi-sensor integration system. For GNSS pseudorange and carrier phase observation data in the urban environment, we offer an innovation-based cycle slip/multipath estimation, detection, and mitigation (I-EDM) method to reduce the influence of multipath effects and cycle slips on location induced by obstruction in urban settings. The method obtains the innovations of GNSS observations with the cluster analysis method. Then the innovations are used to detect the cycle slips and multipath. Compared with the residual-based method, the innovation-based method avoids the residual overfitting caused by the least square method, resulting in better detection of outliers within the GNSS observations. The vehicle tests carried out in urban settings verify the proposed approach. Experimental results indicate that the accuracy of 0.23m, 0.11m, and 0.31m in the east, north and up components can be achieved by the GNSS/INS/Vision tightly coupled system with the I-EDM method, which has a maximum of 21.6 method.

Although the global navigation satellite system (GNSS) and inertial navigation system (INS) integration framework has many advantages compared to the standalone GNSS and INS systems, its accuracy can be severely degraded in urban canyon areas, due to non-line-of-sight (NLOS) signals. Detection and suppression of NLOS signals are the key to improving the location accuracy of urban areas. Therefore, this paper proposes a method for NLOS detection using the improved region growing algorithm on sky-view images captured by a fish-eye camera. After obtaining NLOS detection results, this paper constructs a weighted model with an adaptive scale factor to suppress the influence of NLOS signals. Experimental results show that the proposed method can effectively improve the performance of NLOS detection and suppression. On this basis, a tightly coupled GNSS/INS integration system based on an extended Kalman filter is developed and tested on the vehicle equipment. Results show that the proposed method outperforms the traditional GNSS single point positioning (SPP) and tightly coupled GNSS SPP/INS integration on positioning accuracy. The root mean square error can be reduced by 33.7% and 24.6% in the north and east directions respectively. It indicates that the tightly coupled GNSS SPP/INS integration aided by the fish-eye camera has certain research value and potential on positioning in urban canyon areas.

Cycle-slip detection is crucial in achieving high-accuracy Global Navigation Satellite System (GNSS) data processing for Low Earth Orbit (LEO) satellites. The detector based on the dynamic force model has emerged as a promising approach to cycle-slip detection, as it is insensitive to the number of visible tracked satellites and insufficient accuracy of a prior orbit. However, the ionospheric delay limits the applications of the method for high-speed LEO satellites under different scenarios. To complete the method, we propose a new com-bination of two test parameters: the phase ionosphere-free (IF) combinations in the second-order time-difference model and the Mel-bourne-Wu & BULL;bbena (MW) combinations. Meanwhile, we highlight the identification of whether the fake-positive cycle slips are derived from the flicker noises and jumps in the receiver quartz-based clock. Considering highly, moderately active and quiet ionospheric activ-ities, comparisons are performed among our proposed two test parameters and the phase second-order time-difference wide-lane (WL) combinations. The results indicate that the detector based on the IF combinations performs optimally under the three levels of iono-spheric activities. The time-difference ionospheric delays vary dramatically even under moderately active and quiet ionospheric activities, which leads to the inferiority of the detector based on the WL combinations to one based on the MW combinations. The performances of detectors based on the IF and MW combinations are further evaluated through real-time kinematic orbital determinations comparisons with the Jet Propulsion Laboratory (JPL) precise science orbits (PSO). After using the IF detector, the mean 3-Dimensional (3D) root-mean-squares (RMS) of the orbit differences have slightly decreased by 12% and 6% for the Gravity Recovery and Climate Experiment (GRACE) A and B satellites, respectively. In the case of the GRACE follow-on (GRACE-FO) satellites, the IF detector has achieved a more than 40% improvement over the MW detector. Further, the use of both detectors simultaneously results in a slight improvement in orbit accuracy.& COPY; 2023 COSPAR. Published by Elsevier B.V. All rights reserved.

Ambiguity resolution (AR) is critical for enhancing the orbit accuracy in precise orbit determination (POD) for low earth orbit (LEO) satellites. While orbit estimation using single-difference (SD) AR has been widely researched, the investigation of orbit estimation using undifferenced (UD) AR for LEO satellites is limited due to time-varying hardware biases at the LEO receiver end. To address this deficiency, we propose an improved UD AR method for LEO satellite orbit determination. The method employs the optimal integer datum ambiguity for a 1-day observation arc, and the random-walk clock model is utilized to transfer the integer ambiguity datum to the other epochs, which enables the arc-wise ambiguities to regain the integer property within the time frame. To validate the effectiveness of our improved method, numerical experiments are conducted. Moreover, we assess the performance of the random-walk clock model for a spaceborne ultra-stable oscillator (USO). The high frequency stability of the USO establishes the satisfactory requirements for this study. Both in the kinematic and dynamic modes, the AR success rates of our improved method are - 2% higher than that of the current SD AR method, and the orbit results indicate a slight improvement. The 3-Dimensional (3D) root-mean-squares (RMS) of the orbit differences between JPL precise science orbits (PSO) and our orbits are reduced by up to 4% and 5% for kinematic and dynamic orbits, respectively. The subtle benefit is also proven by K-band ranging (KBR) system validations. In the case of the Gravity Recovery and Climate Experiment Follow-on (GFO) satellites, the results indicate that the effect of our improved UD AR can be equivalent to that of SD AR. Our improved UD AR method can be a good alternative for its superior ability to generate hardware delay at the LEO receiver end. & COPY; 2023 COSPAR. Published by Elsevier B.V. All rights reserved.

北斗三号卫星导航系统(BeiDou-3 navigation satellite system,BDS-3)全球组网工作全面建成,标志着BDS-3迈入全球定位、导航和授时服务的新时代.为了全面比较BDS-3系统与其余全球导航卫星系统(global navigation satellite system,GNSS)非组合精密单点定位(precise point positioning,PPP)性能,重点分析不同分析中心BDS-3精密轨道和钟差产品的一致性、BDS-3/GNSS卫星可用性、BDS-3/GNSS单系统及多系统融合PPP定位性能.结果表明,基于5个分析中心的精密轨道和钟差产品,BDS-3静态PPP三维均方根误差约为2.31～4.00 cm,其单系统收敛时间明显慢于其余GNSS系统,GPS系统的加入对BDS-3/GNSS双系统融合PPP改善效果最为明显,且四系统融合能够有效地缩短收敛时间,并提高动态PPP定位精度.随着BDS-3系统的发展以及轨道和钟差产品的进一步完善,BDS-3同样具备其余GNSS系统提供优质导航定位服务的潜力.

Accurate, continuous and reliable positioning is required in autonomous driving. The precise point positioning (PPP) technique, which can provide a global accurate positioning service using a single global navigation satellite system (GNSS) receiver, has attracted much attention. Nevertheless, due to the cycle slips and multipath effects in the GNSS signal, the performance of PPP is severely degraded in urban areas, which has a negative effect on the PPP/inertial navigation system (INS)/vision integrated navigation. Moreover, the carrier phase observations with un-modeled multipath cause false detection of small cycle slips and lead to deviation in the state variable estimation in PPP. Therefore, an effective cycle slip/multipath estimation, detection and mitigation (EDM) method is proposed. A clustering method is used to separate the cycle slips and multipath from the carrier phase observations aided by visual inertial odometry (VIO) positioning results. The influence of the carrier phase multipath on state variable estimation is reduced by adjusting the stochastic ambiguity model in the Kalman filter. The proposed EDM method is validated by vehicle experiments conducted in urban and freeway areas. Experimental results demonstrate that 0.2% cycle slip detection error is achieved by our method. Besides, the multipath estimation accuracy of EDM improves by more than 50% compared with the geometry-based (GB) method. Regarding positioning accuracy, the EDM method has a maximum of 72.2% and 63.2% improvement compared to traditional geometry-free (GF) and GB methods.