Merging Process Research Articles

A two-step merging strategy for incorporating multi-source precipitation products and gauge observations using machine learning classification and regression over China

Abstract. Although many multi-source precipitation products (MSPs) with high spatiotemporal resolution have been extensively used in water cycle research, they are still subject to various biases, including false alarm and missed bias. Precipitation merging technology is an effective means to alleviate this uncertainty. However, how to efficiently improve precipitation detection efficiency and precipitation intensity simultaneously is a problem worth exploring. This study presents a two-step merging strategy based on machine learning (ML) algorithms, including gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), and random forest (RF). It incorporates six state-of-the-art MSPs (GSMaP, IMERG, PERSIANN-CDR, CMORPH, CHIRPS, and ERA5-Land) and rain gauges to improve the accuracy of precipitation identification and estimation from 2000 to 2017 over China. Multiple environment variables and spatial autocorrelation are combined in the merging process. The strategy first employs classification models to identify wet and dry days and then combines regression models to predict precipitation amounts based on classified wet days. The merged results are compared with traditional methods, including multiple linear regression (MLR), ML regression models, and gauge-based Kriging interpolation. A total of 1680 (70 %) rain gauges are randomly chosen for model training and 692 (30 %) for performance evaluation. The results show that (1) the multi-source merged precipitation products (MSMPs) outperformed all original MSPs in terms of statistical and categorical metrics, which substantially alleviates the temporal and spatial biases. The modified Kling–Gupta efficiency (KGE), critical success index (CSI), and Heidke Skill Score (HSS) of original MSPs are improved by 15 %–85 %, 17 %–155 %, and 21 %–166 %, respectively. (2) The spatial autocorrelation plays a significant role in precipitation merging, which considerably improves the model accuracy. (3) The performance of MSMPs obtained by the proposed method is superior to MLR, Kriging interpolation, and ML regression models. The XGBoost algorithm is recommended more for large-scale data merging owing to its high computational efficiency. (4) The two-step merging strategy performs better when higher-density gauges are used to model training. However, it has strong robustness and can also obtain better performance than original MSPs even when the gauge number is reduced to 10 % (237). This study provides an accurate and reliable method to improve precipitation detection accuracy under complex climatic and topographic conditions. It could be applied to other areas well if rain gauges are available.

Effect of Collision Axes Offset of the Plasmoid in the Collisional Merging Process of FRC Plasma

Effect of Collision Axes Offset of the Plasmoid in the Collisional Merging Process of FRC Plasma

Automatic shrimp counting method using local images and lightweight YOLOv4

Shrimp counting is a fundamental operation for biomass estimation in shrimp culture. It is also vital for achieving reasonable feeding and improving breeding efficiency. The application of computer vision technology in the counting of aquatic products is nondestructive and highly efficient. However, owing to the small size, transparent body, and complex background of shrimp in the actual culture environment, existing methods cannot ensure the lightweight of the model applied while satisfying the counting accuracy, and it is difficult to achieve accurate real-time shrimp counting. Therefore, an automatic shrimp counting method using local images and lightweight YOLOv4 (Light-YOLOv4) was proposed in this study. Multiple local shrimp images were randomly cropped from the original top-view images previously collected using image processing technologies to construct a counting dataset. Subsequently, a local shrimp counting model based on Light-YOLOv4 is constructed and trained using transfer learning. Based on the trained model, the number of shrimp in each local shrimp image was predicted. The number of shrimp in the original shrimp image was obtained through a merging process, and the number of shrimp in the culture area was determined using the frame average method. The method was tested on a real shrimp dataset, and the Light-YOLOv4 local shrimp counting model achieved a counting precision of 92.12%, recall of 94.21%, F1 value of 93.15%, and mean average precision of 93.16%. Compared with other counting models, the proposed method exhibits a better comprehensive performance in terms of the counting accuracy, model size, and detection speed. Furthermore, when shrimp were counted within the entire culture area, the results were consistent with the true values.

Physics of ULIRGs with MUSE and ALMA: The PUMA project

Context. A classical scenario suggests that ultra-luminous infrared galaxies (ULIRGs) transform colliding spiral galaxies into a spheroid-dominated early-type galaxy. Recent high-resolution simulations have instead shown that, under some circumstances, rotation disks can be preserved during the merging process or rapidly regrown after coalescence. Our goal is to analyse in detail the ionised gas kinematics in a sample of ULIRGs to infer the incidence of gas rotational dynamics in late-stage interacting galaxies and merger remnants. Aims. We analysed integral field spectrograph MUSE data of a sample of 20 nearby (z < 0.165) ULIRGs (with 29 individual nuclei) as part of the Physics of ULIRGs with MUSE and ALMA (PUMA) project. We used multi-Gaussian fitting techniques to identify gaseous disk motions and the 3D-Barolo tool to model them. Methods. We found that 27% (8 out of 29) individual nuclei are associated with kiloparsec-scale disk-like gas motions. The rest of the sample displays a plethora of gas kinematics, dominated by winds and merger-induced flows, which makes the detection of rotation signatures difficult. On the other hand, the incidence of stellar disk-like motions is ∼2 times larger than gaseous disks, as the former are probably less affected by winds and streams. The eight galaxies with a gaseous disk present relatively high intrinsic gas velocity dispersion (σ0 ∈ [30 − 85] km s−1), rotationally supported motions (with gas rotation velocity over velocity dispersion vrot/σ0 ∼ 1 − 8), and dynamical masses in the range (2 − 7)×1010 M⊙. By combining our results with those of local and high-z disk galaxies (up to z ∼ 2) from the literature, we found a significant correlation between σ0 and the offset from the main sequence (δMS), after correcting for their evolutionary trends. Results. Our results confirm the presence of kiloparsec-scale rotating disks in interacting galaxies and merger remnants in the PUMA sample, with an incidence going from 27% (gas) to ≲50% (stars). Their gas σ0 is up to a factor of ∼4 higher than in local normal main sequence galaxies, similar to high-z starbursts as presented in the literature; this suggests that interactions and mergers enhance the star formation rate while simultaneously increasing the velocity dispersion in the interstellar medium.

An Unsupervised Clustering Method for Processing Roadside LiDAR Data With Improved Computational Efficiency

In transportation, LiDAR has been primarily used in autonomous vehicles to assist self-driving until recently when people realized it could also be installed at the roadside to support connected vehicles and infrastructure systems. Unlike onboard LiDAR sensors used in autonomous vehicles, roadside applications must perform complete background filtering and clustering as well as tracking real-time traffic movements within the detection zone. This paper presents an unsupervised clustering method for roadside or infrastructure-based LiDAR applications. It first converts 3D LiDAR data points into 2D so that only target points (after background filtering) will be saved in the channel-azimuth 2D structure; then, a method combining the region growing algorithm and counted component labeling is used to perform clustering. Lastly, a merging process is conducted to enhance the connected component labeling method for better outcomes. Experimental studies demonstrate that the proposed method could reach 0.011s per frame (10 Hz sensor rotation frequency) in clustering while maintaining high accuracy.

Solitons in the dark: First approach to non-linear structure formation with fuzzy dark matter

We present the results of a full cosmological simulation with the new codeSCALAR, where dark matter is in the form of fuzzy dark matter (FDM), described by a light scalar field with a mass ofmB = 2.5 × 10−22eV and evolving according to the Schrödinger-Poisson system of equations. In comoving units, the simulation volume is 2.5 h−1Mpc on a side, with a resolution of 20 h−1pc at the highest refinement level. While the resulting large-scale resolution prevents us from studying the general properties of the FDM structure formation, the extremely high small-scale resolution allows a detailed analysis of the formation and evolution of central solitonic cores, which are found to leave their imprints on dark matter density profiles, resulting in shallower central densities, and on rotation curves, producing an additional circular velocity peak at small radii from the centre. Despite the limitations on the large-scale resolution, we find that the suppression of structures due to the quantum nature of the scalar field reveals indications of a shallower halo mass function in the low-mass end compared to the case of a ΛCDM simulation, in which dark matter is expected to cluster at all mass scales even if it was evolved with the same initial conditions as used for FDM. Furthermore, we verify the scaling relations characterising the solution to the Schrödinger–Poisson system for both isolated and merging haloes, and we find that they are preserved by merging processes. We characterise each FDM halo in terms of the dimensionless quantity Ξ ∝ Ehalo/Mhalo3, and we show that the core mass is tightly linked to the halo mass by the core–halo mass relationMcore/Mhalo ∝ Ξ1/3. We also show that the core surface density of the simulated FDM haloes does not follow the scaling with the core radius, as observed for dwarf galaxies. This is a challenge for the FDM model as the sole explanation of core formation.

3D Morphology of Open Clusters in the Solar Neighborhood with Gaia EDR 3. II. Hierarchical Star Formation Revealed by Spatial and Kinematic Substructures

We identify members of 65 open clusters in the solar neighborhood using the machine-learning algorithm StarGO based on Gaia EDR3 data. After adding members of 20 clusters from previous studies we obtain 85 clusters, and study their morphology and kinematics. We classify the substructures outside the tidal radius into four categories: filamentary (f1) and fractal (f2) for clusters <100 Myr, and halo (h) and tidal tail (t) for clusters >100 Myr. The kinematical substructures of f1-type clusters are elongated; these resemble the disrupted cluster Group X. Kinematic tails are distinct in t-type clusters, especially Pleiades. We identify 29 hierarchical groups in four young regions (Alessi 20, IC 348, LP 2373, LP 2442); 10 among these are new. The hierarchical groups form filament networks. Two regions (Alessi 20, LP 2373) exhibit global orthogonal expansion (stellar motion perpendicular to the filament), which might cause complete dispersal. Infalling-like flows (stellar motion along the filament) are found in UBC 31 and related hierarchical groups in the IC 348 region. Stellar groups in the LP 2442 region (LP 2442 gp 1–5) are spatially well mixed but kinematically coherent. A merging process might be ongoing in the LP 2442 subgroups. For younger systems (≲30 Myr), the mean axis ratio, cluster mass, and half-mass–radius tend to increase with age values. These correlations between structural parameters may imply two dynamical processes occurring in the hierarchical formation scenario in young stellar groups: (1) filament dissolution and (2) subgroup mergers.

Atomic-molecular effects in geophysical hydrodynamics

To calculate the dynamics and structure of flows, a system of fundamental equations of fluid mechanics with equations of state for the Gibbs potential and density of an inhomogeneous medium is applied. The complete solution of the system describes ligaments, waves, vortices, jets, wakes, and other types of flows. Calculations of flow patterns around obstacles are consistent with the experiment. Observations of the processes of merging a freely falling drop with a target fluid revealed that the finest components are formed during the direct generation of ligaments by atomic-molecular processes. The involvement of a scaled and parametrically invariant system of fundamental equations permits the study of unsteady energetic flows and more accurately describes their dynamics and structure in the whole range of scales from microscopic to global.

A Rapid Transition from Subversion to Git: Time, Space, Branching, Merging, Offline Commits &amp; Offline builds and Repository Aspects

Background: Software development is the transition from centralized to decentralized version control systems. This transition is driven by the limited features of centralized version control systems in terms of branching, merging, time, space, offline commits &amp; builds and repository aspects. Transition from Subversion; a centralized version control system, to Git; a decentralized version control system has been focused in a limited way. Objective: In this work transition process from Subversion Version Control System (VCS) to Git VCS has been investigated in terms of time, space, branching, merging and repository aspects from the software developer’s point of view; working individually or in a large team over a large and complex software having a legacy of many decades. Experimentation was conducted in SRLC Software Research Lab, Chicago, USA. Methods: Various scripts have been developed and executed for version control using Git and performed over a few legacy software. Results: Results show that branching in Git and Subversion has a difference of about 39 times, i.e. branching operation of Git is about 39 times faster than Subversion. Merging in the case of Git is trivial and automatic, while Subversion needs a manual process of merging, which is error prone. Using an example of Mozilla with fsfs backend, it is observed that disk space can be saved up to 30 times in Git compared to Subversion. By taking a typical example of a large sized project it is observed that Git requires almost half of the revisions compared to Subversion, further with fsfs backend a project having ten years of history with 240,000 commits needs 240 directories in case of Subversion while Git requires only 2 directories. Using offline commits and offline builds of Git, it is observed that in Git whitespace changes, in contrast to significant business logic changes, can be staged in one commit only. These are not possible in Subversion, which requires a complicated system of diffing to temporary files. It is also observed that Git provides offline commit facility, i.e. in case if for some reason, remote repository is unavailable due to disaster or network failure, then still developers can commit their offline code and execute the offline build. Conclusion: However, no previous study was found that focused on how the choice actually affects software developers and it forms the motivation for the present work. In this work, a list of how the choice between Git and Subversion affects software developers is worked out. Although software developers in many aspects are unaffected by the choice, few interesting findings were achieved. One of the most interesting findings of the proposed work is that software developers seem to publish their code to the main repository more often in Git than in Subversion. It is also found that the majority of the software developers perform at least two commits per push, which means that Git repositories will contain a lot more saved points in history than Subversion repositories.

Particle-in-cell modeling of plasma jet merging in the large-Hall-parameter regime

The merging process of magnetized plasma jets with parameters relevant to the plasma-jet-driven magneto-inertial fusion (PJMIF) design and the plasma liner experiment (PLX) is modeled by fully kinetic particle-in-cell (PIC) simulations in one and two spatial dimensions. The modified two-stream instability is identified to be the main mechanism responsible for stopping the plasma jets and preventing species interpenetration. The electron and ion Hall parameters of the merged plasma are greater than unity, and the plasma β is close to unity, which is the desired characteristic of planned experiments at PLX. Our 2D PIC simulations validate the results of the radiation magneto-hydrodynamics code FLASH, which will be the primary tool for modeling various stages of future PJMIF experiments.

Formation of small-scale vortices in the core of a large merged vortex

The merger of two surface quasi-geostrophic vortices is examined in detail. As the two vortices collapse towards each other in the merging process, they trap their external fronts between them; these fronts are inserted into the final merged vortex, where they form a central, nearly parallel, sheared velocity strip, sensitive to barotropic instability. As a result, this strip breaks up into an alley of small vortices. Subsequently, these small vortices may undergo merger and grow in size in the core of the large merged vortex. The number of small trapped vortices decreases correspondingly. Finally, a single or two small vortices remain. These processes are analysed using a numerical model of the surface quasi-geostrophic equations. The sensitivity of this process to the initial vortex characteristics is explored. A parallel is drawn between this problem and the instability of a rectilinear strip of temperature with a central gap. The application of this problem to the Ocean is discussed.

Merging Sequence Optimization Based on Reverse Auction Theory and Merging Strategy with Active Trajectory Adjustment of Heterogeneous Vehicles

This paper investigates the optimized merging sequence (MS) and on-ramp merging strategy in mixed traffic with vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I). To this end, a cooperative merging sequence optimization method is first proposed based on the reverse auction. In the method, a hybrid optimized computing structure is proposed to provide the foundation for connected and human-driving vehicles (CHVs) and connected and automated vehicles (CAVs) to obtain the MS more efficiently. And a cooperative merging strategy based on all cooperative merging vehicles under mixed traffic conditions is proposed. In particular, the downstream vehicles in the strategy can change their original velocities to actively participate in the cooperative merging process according to the merging requirements of on-ramp vehicles. And the vehicles in this strategy are all subject to state constraints to avoid the adverse effects of cooperative merging behavior on the following traffic on the main road. Results of numerical experiments illustrate that the merging sequence optimization method can reduce the time to obtain the optimal MS, and the increased computational efficiency is affected by CAV penetration. In addition, in mixed traffic conditions, the cooperative merging strategy can reduce fuel consumption and the time required for merging.

A puzzle solved after two decades: SN 2002gh among the brightest of superluminous supernovae

ABSTRACT We present optical photometry and spectroscopy of the superluminous SN 2002gh from maximum light to +204 d, obtained as part of the Carnegie Type II Supernova (CATS) project. SN 2002gh is among the most luminous discovered supernovae ever, yet it remained unnoticed for nearly two decades. Using Dark Energy Camera archival images we identify the potential supernova (SN) host galaxy as a faint dwarf galaxy, presumably having low metallicity, and in an apparent merging process with other nearby dwarf galaxies. We show that SN 2002gh is among the brightest hydrogen-poor SLSNe with MV = −22.40 ± 0.02, with an estimated peak bolometric luminosity of 2.6 ± 0.1 × 1044 erg s−1. We discount the decay of radioactive nickel as the main SN power mechanism, and assuming that the SN is powered by the spin-down of a magnetar we obtain two alternative solutions. The first case, is characterized by significant magnetar power leakage, and Mej between 0.6 and 3.2 M⊙, Pspin = 3.2 ms, and B = 5 × 1013 G. The second case does not require power leakage, resulting in a huge ejecta mass of about 30 M⊙, a fast spin period of Pspin ∼ 1 ms, and B ∼ 1.6 × 1014 G. We estimate a zero-age main-sequence mass between 14 and 25 M⊙ for the first case and of about 135 M⊙ for the second case. The latter case would place the SN progenitor among the most massive stars observed to explode as an SN.

Visualization of Macrophase Separation and Transformation in Immiscible Polymer Blends

Visualization of Macrophase Separation and Transformation in Immiscible Polymer Blends

A new three-level mesh method to accelerate the structural topology optimization

To accelerate the structural topology optimization, this paper proposes an adaptive three-level mesh method (ATMM) to reduce the computational costs without loss of accuracy. The ATMM divides the elements into three levels: fine elements, middle elements and coarse elements. Topology optimization is initially performed on the uniform fine meshes, when adjacent fine elements change into fully voids or solids during optimization, they will merge into middle elements, and such merging processes are the same as those between middle elements and coarse elements. Meanwhile, when merged middle elements or coarse elements become gray, they will return to the fine elements. To handle the incompatibility of adjacent elements in different levels, the hybrid-order serendipity elements are adopted. This paper proposes a new nodal numbering scheme to assemble the global stiffness matrix, and a new sensitivity filter scheme is discussed to avoid the numerical instability. Additionally, ATMM can obtain better acceleration and convergence, when combining with the existing gray-scale suppression technology. Lastly, four examples are provided to verify the proposed method, optimization results are consistent with those obtained from the uniform fine meshes, but with greatly reduced computational costs. In the four numerical examples, the time consumed of ATMM in each iteration is only about 30%∼69% compared to that of the uniform fine mesh, and the total iteration numbers can be reduced by 18%∼62%.

Traffic dynamics around freeway merging area with mixed conventional vehicles and connected and autonomous vehicles

The mixed traffic of conventional vehicles (CVs) and connected and autonomous vehicles (CAVs) will dominate a long period in the popularization of CAVs and its dynamics are still not clear, especially in some complicated traffic scenarios. In this paper, an improved cellular automaton model was applied to investigate the CVs–CAVs mixed traffic dynamics around freeway merging area. The lane-changing rules were improved to consider the difference between CAVs and CVs and also the variation of drivers’ aggressive driving during the merging process. The simulation has been conducted with different traffic demands and percentages of CAVs while the phase diagram, spatial-temporal diagram, characteristics of traffic volume, velocity profiles, stability and traffic conflict around the merging area were obtained. The results show that CAVs contribute to improving the traffic operation of freeway merging area in most conditions. However, more CAVs in the merging area without proper management will have a negative impact on traffic if the traffic demand for the main road is great. For example, the stability can decrease and particularly, the on-ramp traffic will be completely blocked when the percentage of CAVs is one. The findings can provide a basis for further research on the CVs–CAVs mixed traffic.

A parameter-free deep embedded clustering method for single-cell RNA-seq data.

Clustering analysis is widely used in single-cell ribonucleic acid (RNA)-sequencing (scRNA-seq) data to discover cell heterogeneity and cell states. While many clustering methods have been developed for scRNA-seq analysis, most of these methods require to provide the number of clusters. However, it is not easy to know the exact number of cell types in advance, and experienced determination is not always reliable. Here, we have developed ADClust, an automatic deep embedding clustering method for scRNA-seq data, which can accurately cluster cells without requiring a predefined number of clusters. Specifically, ADClust first obtains low-dimensional representation through pre-trained autoencoder and uses the representations to cluster cells into initial micro-clusters. The clusters are then compared in between by a statistical test, and similar micro-clusters are merged into larger clusters. According to the clustering, cell representations are updated so that each cell will be pulled toward centers of its assigned cluster and similar clusters, while cells are separated to keep distances between clusters. This is accomplished through jointly optimizing the carefully designed clustering and autoencoder loss functions. This merging process continues until convergence. ADClust was tested on 11 real scRNA-seq datasets and was shown to outperform existing methods in terms of both clustering performance and the accuracy on the number of the determined clusters. More importantly, our model provides high speed and scalability for large datasets.

A Novel Image Denoising and Segmentation Using Machine Learning with SRM Strategy

Objectives: In the last decade, artificial intelligence (AI) and machine learning (ML) have a significant impact on image analysis and segmentation. The most demanding aspect in digital image processing is efficient segmentation to extract the desired features or recognize the hidden patterns of the digital images, which are one of the major challenges. Even in existing segmentation techniques, due to under segmentation issues, particularly in remote or arial image analysis, necessary object features and residual details are not properly segmented. Therefore, the study presents a new method for image segmentation using machine learning, which addresses these limitations. Methods: This article presents a novel approach for arial image denoising and segmentation using unsupervised learning. The Discrete Wavelet Transform (DWT) is primarily used as a pre-processing tool, while edge details of the decomposed image are automatically preserved with the help of machine understanding. The use of a fast Non-Local (NL)-means filter provides a better visual effect for low-frequency image features, and the filtered image is partitioned into the number of clusters by Minimum- Spanning Tree (MST) clustering algorithm. Further, Statistical Region Merging (SRM) is used to eliminate the unwanted regions of the clustered image to give meaningful image details. Finally, the segmented image is projected at the wavelet restoration level. Here, the standard images are extracted from the SIPI database, and the system does not require training prototypes but works in an ”unsupervised” way. Findings: The proposed system segment 1014 regions in the first phase and the segmented image features efficiently recognize the hidden patterns of the arial image (7.1.06). However, to track the required object features, statistical Region Merging (SRM) is utilized, a drastic variation in the merging process eliminates the annoying details and brings effective outcomes, which involves 29 essential features with retained residual details at level-2 decomposition and is compared with leading segmentation techniques. The denoising performance of the suggested study provides a 21.73 Peak Signal to Noise Ratio (PSNR) value in 17.96 elapsed time. simulation results have proven to produce better segmentation and denoising with less iteration time. A qualitative evaluation through visual examination justifies the proposed study and is superior to some of the popular methods. Novelty/Applications: The utilization of wavelet transform to preserve edge features during segmentation is one of the key features of this method, and it significantly overcomes the under segmentation patterns. The combination of minimum-spanning tree clustering (MST) and statistical region merging (SRM) is the major strategic step in segmenting and detecting the essential object features of arial images. Keywords: DWT; Fast NLmeans filter; Wavelet denoising; MST; SRM

Culture Acculturation in Architecture of Masjid Kasimuddin at Bulungan, North Borneo

Acculturation is the process of mixing or merging two different cultures into one new culture. One example of cultural acculturation is found in the architecture of Kasimuddin mosque located in Tanjung Palas, Bulungan Regency, North Kalimantan has a unique architectural form. This mosque is often known as the sultan's mosque because this mosque was built during the Bulungan Sultanate. The existence of this mosque is included in the Bulungan government's cultural tourism project plan. In addition to having a local cultural character, the architecture of this mosque is also influenced by other cultures. This research aims to examine cultural acculturation in the architecture of Kasimuddin Mosque. In addition, it is expected that this writing can be one of the references for the implementation of the next mosque restoration. The research method used is qualitative method. With the form of data collection through field surveys, interviews, and some references from journals. This research was conducted on the Kasimuddin Mosque building. This mosque became one of the artifacts of the Bulungan sultanate and currently still serves as a place of worship as well as a cultural object. Various cultures influence the architecture of the mosque, including Javanese, Sumatran, Betawi, Malay, and European cultures. The occurrence of cultural acculturation because it is located in a coastal area where the area becomes a meeting place between nations. So it is not surprising that in the Bulungan area there is acculturation of different cultures. Kasimuddin Mosque is one example of the result of cultural acculturation. As a cultural heritage building is expected to always be maintained, and considered both by the local government and the local community.

Cooperative Ramp Merging Design and Field Implementation: A Digital Twin Approach Based on Vehicle-to-Cloud Communication

Ramp merging is considered as one of the most difficult driving scenarios due to the chaotic nature in both longitudinal and lateral driver behaviors (namely lack of effective coordination) in the merging area. In this study, we have designed a cooperative ramp merging system for connected vehicles, allowing merging vehicles to cooperate with others prior to arriving at the merging zone. Different from most of the existing studies that utilize dedicated short-range communication, we adopt a Digital Twin approach based on vehicle-to-cloud communication. On-board devices upload the data to the cloud server through the 4G/LTE cellular network. The server creates Digital Twins of vehicles and drivers whose parameters are synchronized in real time with their counterparts in the physical world, processes the data with the proposed models in the digital world, and sends advisory information back to the vehicles and drivers in the physical world. A real-world field implementation has been conducted in Riverside, California, with three passenger vehicles. The results show the proposed system addresses the issues of safety and environmental sustainability with an acceptable communication delay, compared to the baseline scenario where no advisory information is provided during the merging process.