Search

Object detection in traffic scenes has attracted considerable attention from both academia and industry recently. Modern detectors achieve excellent performance under a simple constrained environment while performing poorly under the actual complex and open traffic environment. Therefore, the capability of adapting to new and unseen domains is a key factor for the large-scale application and proliferation of detectors in autonomous driving. To this end, this paper proposes a novel category-induced coarse-to-fine domain adaptation approach (C2FDA) for cross-domain object detection, which consists of three pivotal components: (1) Attention-induced coarse-grained alignment module (ACGA), which strengthens the distribution alignment across disparate domains within the foreground features in category-agnostic way by the minimax optimization between the domain classifier and the backbone feature extractor; (2) Attention-induced feature selection module, which assists the model to emphasize the crucial foreground features and enables the ACGA to focus on the relevant and discriminative foreground features, without being affected by the distribution of inconsequential background features; (3) Category-induced fine-grained alignment module (CFGA), which reduces the domain shift in category-aware way by minimizing the distance of centroids with the same category from different domains and maximizing that of centroids with disparate categories. We evaluate the performance of our approach in various source/target domain pairs and comprehensive results demonstrate that C2FDA significantly outperforms the state-of-the-art on multiple domain adaptation scenarios, i.e., the synthetic-to-real adaptation, the weather adaptation, and the cross camera adaptation.

Background To evaluate the root anatomy, root canal morphology and the anatomical relationship between the roots and maxillary sinus of maxillary second premolars by CBCT in a western Chinese population. Methods A total of 1118 CBCT scans of the maxillary second premolars were collected from West China Hospital of Stomatology, Sichuan University. Information below were measured on axial, coronal and sagittal sections, recorded and evaluated properly: the number of roots and canals, the morphology of canal system classified by Vertucci standard, the inter-orifice distance of canal orifices, the curvature of each canal and the distance from root tip to maxillary sinus floor. Results Among the 1118 teeth, 94.2% (1053) are single-rooted and 55.1% (616) have one canal. Type I (55.1%) is the commonest root canal morphology followed by Type II (31.9%). The mean inter-orifice distance (IOD) for multi-canal teeth ranging from 2.72 ± 0.32 to 3.41 ± 0.11 mm. Of 1622 canals, 38.8% (630) curvature are mesiodistal and 30.9% (501) are straight canals. The distance from root tip to maxillary sinus floor increased with age and the mean distance of single-rooted ones is 2.47 ± 3.45 mm. Conclusions All kinds of canal morphology category can be detected in maxillary second premolars. The IOD might be a predictable factor for root canal morphology. Roots of maxillary second premolars are related to maxillary sinus which should be treated carefully.

Intelligent and connected vehicles have leveraged edge computing paradigm to enhance their environment comprehension and behavior planning capabilities. As the quantity of intelligent vehicles and the demand for edge computing are increasing rapidly, it becomes critical to efficiently orchestrate the communication and computation resources on edge clouds. Existing methods usually perform resource allocation in a fairly effective but still reactive manner, which is subject to the capacity of nearby edge clouds. To deal with the contradiction between the spatiotemporally varying demands for edge computing and the fixed edge cloud capacity, we proactively balance the edge computing demands across edge clouds by appropriate route planning. In this paper, route planning and resource allocation are jointly optimized to enhance intelligent driving. We propose a multi-scale decentralized optimization method to deal with the curse of dimensionality. In large-scale optimization, backpressure algorithm is used to conduct route planning and load balancing across edge clouds. In small-scale optimization, game-theoretic multi-agent learning is exploited to perform regional resource allocation. The experimental results show that the proposed algorithm outperforms the baseline algorithms which optimize route planning and resource allocation separately.

This paper introduces the sea-launch technology of a cryogenic liquid-fueled medium-lift rocket. It first reviews the current state of sea launch technology, and then gives a brief introduction of China’s New Generation Medium-lift Launch Vehicle (NGMLV). The innovations in the NGMVL, such as responsive test and launch control, a H3 launch model, and unmanned operations, provide convenience for sea launches. Based on these innovations, this paper proposes a sea launch scheme, including the system configuration, test and launch processes, and an improved adaptive design for the rocket. Then, the launch platform is discussed in detail, which integrates the functions of sea transportation, assembly and test, as well as technical and launch areas. The layout and function divisions, fluid filling, gas supply and distribution systems, and lossless storage technology of LH2 are described in order. This breakthrough in sea launch technology will enable China to launch medium and large satellites and constellations ‘both on land and sea’, especially into low-inclination Low-Earth Orbits (LEOs), and it allows China to remain competitive in the fast-paced space industry.

This letter theoretically formulates the coded cooperative data exchange (CCDE) problem in multi-channel multi-hop wireless networks, where a set of nodes initially hold a subset of packets and wish to retrieve all desired packets via multi-hop wireless communication. At first, we propose the model of tentative transmission, with which we can jointly consider resource allocation, transmission scheduling, and transmitted packet scheduling. Then, based on the time expanded graph, we convert the CCDE problem in multiple time slots into a single source and multiple sinks network (SSMN) within a time slot. Finally, the relationship between the required time slots and the number of channels is theoretically formulated based on the constructed SSMN. Experiments demonstrate that the gain due to the increasing number of channels flattens.

Intelligent driving plays a role in significantly improving the safety and efficiency of transportation systems. As the onboard capabilities of perception, comprehension, and decision making are limited, vehicles can employ the edge computing infrastructure of the smart road to enhance their intelligence. Therefore, the smart road is considered an intelligent Internet of Things system. It provides vehicles with not only the road space in the transportation domain, but also the communication, sensing, and computing resources in the information domain to improve the composite quality of intelligent driving. However, the resources in the information and transportation domains are complicatedly coupled, and the orchestration of these cross-domain resources is confronted with the huge state-action space, which cannot be solved in a real-time manner. In this article, we investigate the fundamental research challenges in cross-domain resource orchestration for the smart road, and design a multi-agent-based framework. Within the framework, each vehicle is associated with an exclusive agent on the edge cloud, and the agents utilize swarm intelligence to jointly optimize the traffic flow and information flow for their respective vehicles. Specifically, a value iteration network is used by agents to learn the routing behavior of vehicles, and a multi-agent deep reinforcement learning method is proposed, enabling agents to cooperatively learn decentralized resource optimization policies. To verify the effectiveness of the proposed framework, a cross-domain resource orchestration prototype is implemented and evaluated.

In this work, flower-like Bi2MoO6 nanoparticles grown on FTO substrates were firstly fabricated using a seed-free hydrothermal method. The Bi2MoO6 nanoflowers exhibited, to the best of our knowledge, higher photoelectrochemical (PEC) performances than other previously reported morphologies. It is generally accepted that the formation of type-I heterostructures is unfavorable for PEC applications. Nevertheless, in this work, we have successfully constructed a novel type-I architecture with numerous electron transport channels. In this unique Bi2MoO6/BiVO4 structure, BiVO4 films were continuously distributed on both the surfaces and the interstices of Bi2MoO6 nanoflowers. Interconnected BiVO4 nanoparticles could intimately contact with FTO substrates and thus constitute the electron transport channels, which could promptly transfer electrons to FTO substrates. Simultaneously, a cocatalyst of g-C3N4 was modified on the surfaces of BiVO4 to capture the photogenerated holes. As a result, the PEC activities of Bi2MoO6/BiVO4 heterostructures were significantly improved due to the enhanced charge carriers separation efficiency. The special design of electron transport channel may provide a universal strategy to address the intrinsic drawbacks of type-I heterostructures.

This article investigates the coded cooperative data exchange (CCDE) problem, where a set of nodes initially hold a subset of packets and wish to retrieve all desired packets via direct wireless communication with neighbors. The CCDE problem in multihop wired networks has seen significant research recently and is proved to be NP-hard. The CCDE problem in multihop wireless networks (MWNs) must additionally consider half-duplex constraint, interference constraint, and channel constraint. Channel assignment brings new challenges to the CCDE problem in MWN, since it is also a well-known NP-hard problem even with one channel. In this article, we study the CCDE problem in MWN with multiple channels. We first construct a path network to evaluate the priority for each possible transmission and then construct a conflict graph (CG). This graph depicts the half-duplex, interference, and channel constraints. After that, a greedy channel assignment algorithm is proposed to obtain the nonconflict transmissions based on the constructed CG and assign a channel for each selected transmission. Finally, based on the selected transmissions, we construct a single-source multicast network exploiting the time expanded network, with which the network encoding and decoding schemes can be computed within polynomial time. Extensive simulations are conducted to demonstrate the efficiency of the proposed algorithm.

Rosacea is a common chronic skin disorder of unknown etiology. While population prevalence rates range 0.2-22% in Europe and North America, prevalence in China is currently undetermined. We conducted a large population-based case-control study to determine the present epidemiological status of rosacea in China, involving 10 095 participants aged 0-100 years (mean age, 35.5 ± 19.1; 50.5% female). A census of rosacea among 15 communities in Changsha in south central China was conducted with skin examination by board-certified dermatologists. Rosacea was observed in 3.48% (95% confidence interval, 3.13-3.85%) of the study population. Subtype distribution was erythematotelangiectatic in 47.6%, papulopustular in 35.0% and phymatous in 17.4%. Family history was noted in 37.8% and ocular symptoms in 31.3%. Associations with rosacea were observed for melasma, hypertension, hyperthyroidism and breast cancer in females (P < 0.05), and also for hyperthyroidism and peptic ulcers in males (P < 0.05). Our results provide baseline information about epidemiological aspects of rosacea in China.

In this article, an optical fiber SPR sensor based on polymer-tipped optical fiber (PTOF) coated with a gold film and a Barium Titanate (BaTiO3) layer has been designed and optimized by finite difference time-domain (FDTD) method. Three SPR dips are observed and the optimal sensitivity of single-mode PTOF SPR sensor coated with an additional BaTiO3 layer reaches 6710 nm/RIU when refractive index (RI) of analyte is 1.333–1.343. A novel measurement method based on sum-value and difference-value of two SPR dips is proposed. Compared with conventional measurement method based on single dip, the measurement based on sum-value can significantly improve the sensitivity and the measurement based on difference-value can effectively reduce the environmental influence.

Recent years have witnessed more attention being paid to 3D object detection, due to its indispensable roles in autonomous driving and intelligent transportation systems. The accuracy of 3D object detection heavily depends on sufficient and robust feature extractors. However, most feature extractors adopt fixed perceptive fields or fixed spatial locations, which could not efficiently model geometric variations or geometric transformations in object scale, pose, and part deformation. To fill this gap, this paper proposes a novel Shape-Invariant Set-Abstraction network (SI-RCNN) for 3D object detection in point clouds, which can extract shape-adaptive features. Specifically, we first propose an Offsets Generation Module to learn a set of offsets, where the shape of the offsets can adapt to the changes of features. Then, trilinear interpolation is applied to assign features to each offset, since it might be empty due to discrete points in the cloud data. Finally, we further exploit the ground-truth box to supervise the offsets. Extensive experiments demonstrate the effectiveness of our proposed SI-RCNN on the competitive KITTI 3D detection benchmark, as compared to several state-of-the-art methods.

Multi-Object Tracking (MOT) which aims to estimate trajectories for objects of interest in videos, is a crucial component in intelligent transportation systems and has gained considerable attention from both academia and industry. However, existing MOT approaches apply simple motion and appearance features, ignoring the spatial-temporal correlations of interactions with other objects. To fill this gap, this paper proposes a novel MOT framework ST-Tracking to extract spatial-temporal features for tracking objects, which is trained in an end-to-end manner. Specifically, a directed graph is first constructed to represent the spatial-temporal correlations, where the node attributes include both geometric and appearance features, and the edges are deliberately defined to embrace location and topology cues. Then, a spatial-temporal graph convolutional module is proposed to aggregate information from both spatial and temporal dimensions, thus extracting discriminative features. Extensive experiments on the challenging KITTI benchmark demonstrate the effectiveness and superior of ST-Tracking as compared with several state-of-the-art methods.

Humans are not robots with predefined path planning rules. Pedestrian trajectory prediction is effected by the physical world's laws, social space and their own personality. In this paper, we represent a novel insight walking rhythm and social relation-potential model (WR-SRPG) to solve these problems. We propose a new way to measure the information of pedestrian walking characteristics and personality characteristics. We design a Social Relation-Potential Grid (SRPG) module to model the complex social interaction between pedestrians. And we integrate multi-head attention mechanism, position encoding and remote estimation points to expand the ability of the model to focus on different locations and model long-term spatiotemporal dependence. The experimental results show that our method surpass state-of-the-art method on ETH&UCY datasets in ADE and FDE, and successfully predicts the complex social behaviors among pedestrians.

3D object detection has attracted a significant amount of attention and interest from both academia and industry due to its indispensable role in understanding 3D environments. By fusing the camera and LiDAR sensors, it is expected to improve both the accuracy and robustness of 3D object detection. However, existing fusion approaches are either limited by cascading processing, or easy to be influenced by the interference information in multi-sensors. To this end, this paper incorporates meta learning to fuse the camera and LiDAR data. Specifically, we first extract meta knowledge from images, and then apply it to generate the parameter weights of a set of convolution kernels, which are further exploited for feature extraction on LiDAR point clouds. Furthermore, we propose a meta fusion network (MF -Net), enabling accurate and robust 3D object detection. The superiority and effectiveness of MF -Net have been demonstrated by extensive experiments on KITTI 3D object detection dataset.

The infrastructure to vehicle (I2V) communication boosts a large number of prevailing vehicular services, which can provide vehicles with external information, storage, and computing power located at both mobile edge server (MES) and remote cloud. However, vehicle distribution is imbalanced due to the spatial inhomogeneity and temporal dynamics. As a consequence, the communication load for MES is imbalanced and vehicles may suffer from poor I2V communications where the MES is overloaded. In this paper, we propose a novel proactively load balancing approach that enables efficient cooperation among MESs, which is referred to as end-to-end load balancer (E2LB). E2LB schedules the cached data among MESs based on the predicted road traffic situation. First, a convolutional neural network (CNN) is applied to efficiently learn the spatio-temporal correlation in order to predict the road traffic situation. Then, we formulate the load balancing problem as a nonlinear programming (NLP) problem and a novel framework based on CNN is adopted to approximate the NLP optimization. Finally, we connect the above neural networks into an end-to-end neural network to jointly optimize the performance, where the input is the historical traffic situation while the output is the balanced scheduling solution. E2LB can guarantee the real-time scheduling, since the calling of a well-trained neural network only requires a small number of simple operations. Experiments on the trajectories of taxis and buses in Beijing demonstrate the efficiency and effectiveness of E2LB.

To identify and understand parathyroid lesions of patients with primary hyperparathyroidism (PHPT) more accurately under ultrasound.This retrospective study involved 423 adult patients with PHPT with a single parathyroid nodule and positive parathyroid ultrasonography between 2018 and 2019. The clinical characteristics of the study patients and histopathologic sections were reviewed.According to the main grayscale echogenicity features of parathyroid nodules, 423 cases were divided into groups: iso-hyperechogenicity solid (61/423), hypoechogenicity solid (304/423), and mixed-echogenicity cyst-solid (58/423) groups. Comparison among the 3 groups showed that the iso-hyperechogenicity group included more asymptomatic patients with PHPT and fewer patients with severe symptoms like bone fractures (P.05). The mixed-echogenicity group showed higher median serum parathyroid hormone (PTH) and serum calcium levels and larger lesion sizes (P.05), and the iso-hyperechogenicity group showed the lowest median serum PTH level. No difference in lesion size was noted between the 2 solid groups, but the median serum PTH level in the hypoechogenicity group was higher than that in the iso-hyperechogenicity group (P.05). According to histopathology, the hypoechogenic area of the samples may contain more functional components (chief cells), whereas the iso-hyperechogenic area has more nonfunctional components (eg, lipocytes and connective tissues).The PHPT nodules distinguished by ultrasound echogenicity features showed different histopathologic components, reflected by different clinical characteristics of the patients with PHPT.

Objective. To explore the influencing factors of medication compliance in patients with recurrent vertebral fractures after percutaneous kyphoplasty (PKP) and the role of family-centered education intervention. Methods. From January 2018 to January 2021, the general disease-related data survey form and medication compliance questionnaire made by our hospital were used to evaluate the scores of 198 patients with recurrent vertebral fractures after PKP in the Department of Orthopedics of our hospital. Single-factor and multiple linear regression analyses were used to explore the influencing factors of medication compliance in patients with recurring vertebral fractures after PKP. From 198 patients, 80 eligible patients were selected for further research. According to a random number table method, they were divided into the control group (n = 40) given only antiosteoporosis drug treatment and care and the experimental group (n = 40) combined with family-centered education intervention. After 12 months of intervention, the two groups were evaluated for their knowledge of osteoporosis, medication compliance, and physical health. Results. Of the 198 patients, only 65 had good medication compliance, 90 had poor medication compliance, and 43 were acceptable. Univariate analysis showed that the influencing factors of medication compliance in patients with recurrent vertebral fractures after PKP include the patient’s education, living style, per capita monthly income, combined other diseases, number of hospitalizations, and time since the last hospitalization ( P < 0.05 ). Multiple linear regression analysis showed that patients with recurring vertebral fractures after PKP with high education, living with spouse or children, combined with other diseases, frequent hospitalizations, and short time from the last hospitalization had higher medication compliance ( P < 0.05 ). After the intervention, the disease knowledge mastery of the experimental group was significantly better than before and after the intervention in the control group ( P < 0.0001 ). After the intervention, the medication compliance and health status of the experimental group were significantly better than those of the control group ( P < 0.05 ). Conclusion. The medication compliance of patients with recurrent vertebral fractures after PKP is generally poor, and medical staff need to take targeted interventions based on the main factors that affect the patients’ medication compliance. Family-centered education intervention is an effective way to improve disease awareness, medication compliance, and health status of patients with recurring vertebral fractures after PKP.

In this work, we demonstrate the fabrication of Mo, W co-doped BiVO4/ZnO nanosphere (Mo, W: BVO/ZnO NS) heterostructures using a simple solution dry-out method that uniformly embeds ZnO NSs in Mo, W: BVO nanocrystals. Photoelectrochemical (PEC) examination confirmed that Mo, W: BVO/ZnO NSs exhibit higher PEC performance than pure Mo, W: BVO, and this improved performance of Mo, W: BVO/ZnO NSs also depends on decreased ZnO NS size. Moreover, a layered ZnO/Mo, W: BVO (ZnO film coated with Mo, W: BVO layer) heterostructure was prepared using a simple physical piled method that exhibited far lower PEC performances than those of Mo, W: BVO/ZnO NS heterostructures. These results clearly revealed that the formation of the Mo, W: BVO/ZnO NS heterostructure could increase the charge carrier density and it possessed a much larger contact area. These improvements were favorable for the transfer and transport of photoexcited charge carriers. This work offers an effective strategy to fabricate heterostructures with effective charge separation by a simple solution dry-out method that is suitable for other heterostructure-engineered photoanodes in solar water-splitting applications.

Nowadays navigation systems are mostly centralized and mainly relying on the traffic conditions of road network, and other influence factors are rarely considered, such as roads topology, domain knowledge and planning autonomy, which is result in the recommended seemingly optimal routes may attract numerous vehicles and then congestion. Besides, the existing region-based methods have deviation on acquiring the traffic conditions of road segments, which lead to the inadequate use of road resource and the waste of drivers’ time.Considering that MEC can provide low latency and quick responses support for intelligent driving, we propose an online planning approach on the road graph via a double rewarded Generalized Value Iteration Network, 2r-GVIN for short. In detail, a large-scale real GPS trajectories are mapped and discretized into traffic sequence pairs anonymously, that is, capturing the traffic status and driving actions as dataset. Then a neural network ConvLSTM which can effectively capture the spatio-temporal properties, is trained to predict the future traffic situation to form “prediction reward”. To enable learning and planning on irregular road graphs straightforwardly, an economical embedding graph convolution operator network GVIN is used. Combining with the model input “current reward”, 2r-GVIN is formed to solve the plan-involved function iteratively. Lastly, we train and evaluate 2r-GVIN on planning problem in road networks, showing that it can achieve road segment-based self-defined navigation with high success rate.

Rising traffic congestion has become a major challenge to urban areas. Route planning and traffic signal timing are widely used methods to improve traffic efficiency. Although these two optimization problems are tightly coupled with each other, they have traditionally been studied separately in existing works. In this paper, we propose a decentralized framework that enables dynamic orchestration of route planning and traffic signal timing to improve traffic efficiency. In this framework, vehicles and traffic lights coordinate to offload traffic optimization tasks to their respective virtual agents on the edge cloud. Furthermore, considering the group rationality of agents, we formulate the joint optimization problem as an evolutionary game. With effective interactions, the agents leverage replicator dynamics to find the Nash equilibria for the evolutionary game, which can coordinate the integrated behavior of multiple vehicles and traffic lights. The simulation results show that the proposed method outperforms the baseline solutions, which optimizes the route planning and signal timing separately.

Urban traffic congestion has a great impact on commute time, energy consumption, and carbon emissions. To deal with traffic congestion, the vehicle cooperative routing method coordinates the routing behavior of vehicles according to dynamic traffic demands. However, city-wide cooperative routing is faced with extremely high communication and computing complexity, which makes it difficult to guarantee real-time performance. To this end, we propose a hierarchical traffic-balanced route planning method for connected vehicles based on edge computing. Specifically, the road network is divided into grids, and an improved back-pressure algorithm is used to guide the inter-grid traffic flow. Furthermore, to balance the intra-grid traffic flow, the vehicle routing is scheduled by evolutionary game method at each intersection. The simulation results show that the algorithm can effectively balance the utilization of road network resources, increase the throughput of the road network and reduce the total traffic time.

The existing end-to-end methods make driving decisions mainly based on the vehicle's own perceived data, which cannot avoid hazards in blind zones. To fill this gap, vehicles should cooperate to construct a comprehensive environment perception by sharing information among each other, equipping each vehicle with see-through ability. While bringing more perceived information, data from other sources may also interfere feature selection and make decision making more difcult. To solve this problem, we propose a dual-attention-based neural network by utilizing two different attention modules. The first module is designed for each source to eliminate redundant features in perception and generate cognitive information for sharing. Since the influences of different cognition on the decision making are different under different circumstances, the second module is used to discriminate the importance of different cognition and focus on the dominant one as needed. Guided by the dual-attention-weighted features, the proposed network extracts the most salient features from the multi-source data, which leads to a signicant reduction of false response in steering angle controlling. Extensive experiments have demonstrated the superior performance of our proposed method, as compared with several state-of-the-arts.

Wireless rechargeable sensor network is a new type of sensor network which has been widely concerned recently. It has the function of charging nodes in network. We use the mobile energy replenishment device to actively move to nodes in the sensor network for charging. Research shows that different charging circuit order will get different charging performance. Therefore, we explore the design of scheduling scheme to achieve better charging performance. In this paper, three new concepts are proposed respectively: the start moment of charging, the energy exhaustion moment and the state of charging to evaluate the scheduling. Based on the above, we define a quadratic performance index as the evaluation of charging performance in sensor network called charging cost of network, the smaller the index is, the better charging performance of the network will be. Through theoretical analysis, firstly, we prove the problem in this paper is NP-complete. Then, considering the charging efficiency of each node to be charged, combined with the quantum behavior particle swarm optimization(QPSO) algorithm, we propose an optimal scheduling scheme for mobile energy replenishment device. Finally, some simulations and theoretical analyses can verify the correctness and effectiveness of the scheme which we proposed.

With the expansion of the mobile communication technology, telecommunication fraud is increasing dramatically which results in the loss of billions of dollars worldwide every year. In recent years, some detection methods utilize data mining and statistical techniques to detect fraud from large amount of subscriber content data, and some methods transform the social relation into a set of topological features, such as degree, k-core etc. However, both content and relation have not been fully explored for identifying fraudsters. In this paper, we propose the Multi-Range Gated Graph Neural Network (MRG-GNN) for learning latent features of social network. Specifically, we first model a social network as a directed graph where vertices with subscriber features represent subscribers and edges with relational features represent activities between them. Then, a novel method based on efficient short walks and node-merging is proposed to structure the convolutions, and graph convolution block captures content information and relation information between users. The multi-range gated readout operation is proposed to aggregate informative features in multiple nodes and automatically learns the representation of user social network. Finally, experiments on a real-world telecommunication network show that our MRG-GNN achieves the state-of-the-art results.

In this paper, an optical fiber SPR sensor based on polymer-tipped optical fiber (PTOF) coated with a gold film and a Barium Titanate (BaTiO3) layer has been designed and analyzed by finite difference time-domain (FDTD) method. Three SPR dips are observed and the sensitivity is proved to be higher than the structure without BaTiO3. A novel measurement method based on sum-value and difference-value of two SPR dips is proposed. Compared with conventional measurement method based on single dip, the measurement based on sum-value can significantly improve the sensitivity and the measurement based on difference-value can effectively reduce the environmental influence.

The challenge in ensuring the reliability of autonomous vehicles is full awareness of the surrounding environment and high-precision steering control. The latest solutions to this challenge include deep learning technologies that provide end-to-end solutions to predict steering angles directly from environmental cognition information with high accuracy. Under the background of 5G technology, edge device has certain computing power, which can reduce the load of on-board computing equipment. In this paper, we present a new distributed perception-decision network model. This model allows the network’s computing tasks to be offloaded to the edge computing devices to reduce the consumption of vehicle-mounted computing devices. The feasibility of the model is verified by experiments. Compared with the existing methods, the model also has a higher accuracy of steering prediction.

The development of Internet of vehicle (IOV) and autopilot technology indicates the coming of smart traffic and automatic unmanned era. The promotion of networking and intelligence not only provides a rich source for urban traffic data, but also provides an efficient and direct way to solve urban traffic problems. With the help of deep learning and reinforcement learning technology, we propose a model to mine the urban traffic rule from the travel history of urban travelers, and utilize it achieving better allocation of traffic resources by providing a traffic guidance service, finally realize the system optimal traffic travel.

This paper proposes a traffic signal cooperative control algorithm based on multi-agent reinforcement learning (MARL), and design a framework of edge computing under traffic signal control scene. By introducing edge computing into the scene of traffic signal cooperative control, it will bring minimal response time and reduce network load. We abstracted the traffic signal control problem into the Markov decision process (MDP). The traffic state is discretized by feature extraction to avoid the curse of dimensionality. We propose a fusion of multi-agent reinforcement learning and coordination mechanisms through collaborative Q-values. The action selection strategy of an intersection depends not only on its own local reward, but also on the impact of other intersections. Different from considering only adjacent intersections, algorithm combines the static distance and dynamic traffic flow, and considers the cooperative relationship between neighbor and non-neighbor nodes. Finally, we show through simulation experiments on SUMO that our algorithm can effectively control traffic signal.

Hepatoid adenocarcinoma of lung (HAL) is a rare and aggressive tumor. The current study reported a new HAL case in the right lower lung with high serum α-fetoprotein (AFP) level in a 71-year-old male patient. After the confirmation of morphology and immunohistochemistry, the patient was diagnosed clinically with HAL and treated with radio-frequency ablation. However, the patient whose disease progressed eventually died 4 months after diagnosis. Whole genome sequencing analysis identified a driver gene mutation in the FAT atypical cadherin 1 gene (FAT1) and the copy number loss. The tumor was microsatellite-stable and tumor mutation burden (TMB) was 1.69 mutations/Mb. PD-L1 expression was negative by IHC. Our finding provide further clues for the molecular basis of HAL and the efficacy of immunotherapy needs to be explored.

The performance of a vehicular ad hoc network (VANET) largely depends on the underlying medium access control (MAC), as it determines the schedule utility of physical resources. However, in existing time-division multiple access (TDMA)-based MAC protocols, a node usually acquires slots based on what each node senses, which is typically the coupling of the control and data plane. This coupling makes the TDMA protocols unable to rapidly and agilely deal with the challenges in VANETs, such as high mobility and dynamic network densities. Inspired by the software-defined network (SDN), we propose a novel SDN-based MAC protocol, named sdnMAC, to handle these challenges. A novel roadside openflow switch (ROFS) is designed as the roadside unit, controlled by the openflow controller. The sdnMAC can be divided into two tiers, the management of ROFSes (MA-ROFS) by the controller and the management of vehicles (MA-VEH) by ROFSes. In MA-ROFS, the controller schedules the cooperative sharing of time slot information among ROFSes. In MA-VEH, each ROFS allocates slots based on this shared information, thus decoupling of the control and data plane. This decoupling provides great rapidness and agility to sdnMAC, thus handling the rapid mobility and varying vehicle densities. Extensive simulations using network simulator NS3 and traffic simulator SUMO are performed. It is shown that the sdnMAC protocol can better meet the requirements of cooperative safety in VANETs.

Background After the promotion of the two-child policy in recent years, the population of children in mainland China was bound to have a rapid growth, which would bring great challenges to public health. A number of cross-sectional studies on the epidemic of childhood asthma in mainland China were recently conducted, and varied prevalences were reported. Thus, knowing the epidemiology of childhood asthma in mainland China is of great necessity. Objective Our study aimed to summarize the pooled prevalence of childhood asthma in mainland China and its time trend, gender difference, regional distribution, and age structure. Methods Studies that reported the prevalence of childhood asthma in mainland China were identified via a systematic data base search through July 1, 2016. Meta-analysis was used to estimate the prevalence of childhood asthma and its subgroups, including gender, age groups, years, and regions. The regional distribution of the prevalence was set by province with the help of a geographic mapping software. The autoregressive integrated moving average model was used to predict the current prevalence of asthma. Results A total of 117 studies published from 1988 to 2014 in mainland China with a total sample size of 2,678,696 were included. The overall current prevalence and lifetime prevalence of childhood asthma was 2.112% (95% confidence interval [CI], 1.977-2.247%) and 2.502% (95% CI, 2.166-2.838%), respectively. The difference of the prevalences between male and female patients was significant: odds ratio 1.54 (95% CI, 1.47-1.62) for the current prevalence and odds ratio 1.61 (95% CI, 1.47-1.77) for the lifetime prevalence. Conclusion The prevalence of childhood asthma in mainland China was low but has been increasing remarkably since 1998. Boys are more likely to have asthma throughout most of their childhood. Preschoolers (3-6 years old) showed a higher prevalence than the other age groups. The current prevalence of childhood asthma probably increased slightly from 2017 to 2019.

Automated driving is coming with enormous potential for safer, more convenient, and more efficient transportation systems. Besides onboard sensing, autonomous vehicles can also access various cloud services such as high definition maps and dynamic path planning through cellular networks to precisely understand the real-time driving environments. However, these automated driving services, which have large content volume, are time-varying, location-dependent, and delay-constrained. Therefore, cellular networks will face the challenge of meeting this extreme performance demand. To cope with the challenge, by leveraging the emerging mobile edge computing technique, in this article, we first propose a two-level edge computing architecture for automated driving services in order to make full use of the intelligence at the wireless edge (i.e., base stations and autonomous vehicles) for coordinated content delivery. We then investigate the research challenges of wireless edge caching and vehicular content sharing. Finally, we propose potential solutions to these challenges and evaluate them using real and synthetic traces. Simulation results demonstrate that the proposed solutions can significantly reduce the backhaul and wireless bottlenecks of cellular networks while ensuring the quality of automated driving services.

Popular navigation services are used by drivers both to plan out routes and to optimally navigate real time road congestion in internet of vehicles (IoV). However, the navigation system (such as GPS navigation system) and apps (such as Waze) may not be possible for each individual user to avoid traffic without creating congestion on the clearer roads, and it might even be that such a recommendation leads to longer aggregate routes. To solve this dispersion, in this paper, we first apply a concept of virtual vehicle in IoV, which is an image of driver and vehicle. Then, we study a setting of non-atomic routing in a network of m parallel links with symmetry of information. While a virtual vehicle knows the cost function associated with links, they are known to the individual virtual vehicles choosing the link. The virtual vehicles adapt the cooperation approach via strategic concession game, trying to minimize the individual and total travel time. How much benefit of travel time by the virtual vehicles cooperating when vehicles follow the cooperation decisions? We study the concession ratio: the ratio between the concession equilibrium obtained from an individual optimum and the social optimum. We find that cooperation approach can reduce the efficiency loss compared to the non-cooperative Nash equilibrium. In particular, in the case of two links with affine cost functions, the concession ratio is at most 3/2. For general non-decrease cost functions, the concession ratio is at most 2. For the strategic concession game, the concession ratio can approach to 1 which is a significant improvement over the unbounded price of anarchy.

Cloud computing is becoming an important solution for providing scalable computing resources via Internet. Because there are tens of thousands of nodes in data center, the probability of server failures is nontrivial. Therefore, it is a critical challenge to guarantee the service reliability. Fault-tolerance strategies, such as checkpoint, are commonly employed. Because of the failure of the edge switches, the checkpoint image may become inaccessible. Therefore, current checkpoint-based fault tolerance method cannot achieve the best effect. In this paper, we propose an optimal checkpoint method with edge switch failure-aware. The edge switch failure-aware checkpoint method includes two algorithms. The first algorithm employs the data center topology and communication characteristic for checkpoint image storage server selection. The second algorithm employs the checkpoint image storage characteristic as well as the data center topology to select the recovery server. Simulation experiments are performed to demonstrate the effectiveness of the proposed method.

In recent years, IoV (Internet of Vehicles) has become one of the most active research fields in network and intelligent transportation system. As an open converged network, IoV plays an important role in solving various driving and traffic problems by advanced information and communications technology. We review the existing notions of IoV from different perspectives. Then, we provide our notion from a network point of view and propose a novel IoV architecture with four layers. Particularly, a novel layer named coordinative computing control layer is separated from the application layer. The novel layer is used for solving the coordinative computing and control problems for human-vehicle-environment. After summarizing the key technologies in IoV architecture, we construct a VV (Virtual Vehicle), which is an integrated image of driver and vehicle in networks. VVs can interact with each other in cyber space by providing traffic service and sharing sensing data coordinately, which can solve the communication bottleneck in physical space. Finally, an extended IoV architecture based on VVs is proposed.

With the rapid progress of cloud computing technology, a growing number of big data application providers begin to deploy applications on virtual machines rented from infrastructure as a service providers. Current infrastructure as a service provider offers diverse purchasing options for the application providers. There are mainly three types of purchasing options: reserved virtual machine, on-demand virtual machine and spot virtual machine. The spot virtual machine is a specific type of virtual machine that employs a dynamic pricing model. Because can be stopped by the infrastructure as a service providers without notice, the spot virtual machine is suitable for large-scale divisible applications, such as big data analysis. Therefore, spot virtual machine is chosen by many big data application providers for its low rental cost per hour. When spot virtual machine is chosen, a major issue faced by the big data application providers is how to minimize the virtual machine rental cost while meet service requirements. Many optimal spot virtual machine purchasing approaches have been presented by the researchers. However, there is a shortage of simulators that enable researchers to evaluate their newly proposed spot virtual machine purchasing approach. To fill this gap, in this paper, we propose SpotCloudSim to support for dynamic virtual machine pricing model simulation. SpotCloudSim provides an extensible interface to help researchers implement new spot virtual machine purchasing approach. In addition, SpotCloudSim can also study the behavior of the newly proposed spot virtual machine purchasing approaches. We demonstrate the capabilities of SpotCloudSim by using three spot virtual machine purchasing approaches. The results indicate the benefits of our proposed simulation system.

Continuum manipulators are particularly suitable for performing tasks in cluttered environments with limited space for maneuvering. While there is progress on continuum grasping of an object in a cluttered space, if the environment is unknown where the target object is occluded and only partially visible, such as in a search and rescue scenario, an open problem is how to determine if it is possible to fetch the object. In this letter, we address this problem of online determining whether a partially occluded object nested in an unknown cluttered space can be fetched by a continuum manipulator based on sensing the surrounding obstacles progressively with a distance sensor attached to its tip, and if so, how to fetch the object autonomously. Our method formulates constraints that can be quickly checked from sensed information to decide if a solution exists for a multisection spatial continuum manipulator to access and grasp an object in a cluttered space. Examples using point clouds of unknown obstacles from RGB-D sensing illustrate the effectiveness of our approach.

Increasing number of vehicles is making congestion problem become more and more deteriorate. This problem could be alleviated by route planning which guides vehicles to routes with small traffic flows to get the shortest travel time. However, the existing route planning algorithms mostly focus on one single vehicle and overlook the coordination among vehicles. If all vehicles follow the same routing recommendation, a large traffic volume will flow into the same route and cause congestion on that route. That makes the routing method ineffective. To resolve this problem, a distributed cooperative routing algorithm (DCR) based on evolutionary game theory is proposed to coordinate vehicles. This method runs on roadside units (RSUs) with combination of edge computing and edge intelligence. A road network is built to evaluate the performance of proposed algorithm. The experiment results show that the proposed DCR algorithm balances the distribution of traffic flow and in the same time makes the total travel time from origin to destination smaller.

Guiding vehicles to their destination is an essential service. Nowadays navigation systems are mainly relying on the traffic conditions of road network, and other influence factors are not taken into account accurately, which is easy to lead to imbalance between the supply and demand of roads, resulting in congestion. In this paper, we introduce an online guiding approach via value iteration network on road graphs, R-VIN for short, which is an end-to-end planning model. In R-VIN, a large-scale real GPS trajectories are mapped via map-matching based road topology, which enables R-VIN to catch the experienced driving knowledge. Then we propose a conversion method from irregular road graphs to regular grid images to formalize the learning model. For a global optimum, ConvLSTM is used to predict the future traffic situation to form "prediction reward" of R-VIN. Combining with "current reward", a double rewarded VIN is used to solve the plan-involved function. Lastly, we train and evaluate R-VIN on planning problem in road networks, showing that R-VIN can achieve segment-based autonomous navigation with high top-k accuracy and less commuting time.

Isolated bidirectional dc/dc converters (IBDDCs) are widely used in the renewable energy systems such as solar energy and wind power as the power interfaces. Dual active bridge (DAB) converters show many advantages compared with other IBDDCs, such as high power density and easy in control implementation. Nowadays, many advanced controls with multiple phase-shift variables such as triple phase-shift (TPS) control have been proposed to improve the efficiency for a wider output power range. In order to ensure the system reliability, the fault diagnosis is essential to identify the faulty devices quickly and avoid the system collapse. In this paper, three common operation modes are analyzed and an open-circuit diagnosis strategy is proposed for DAB converters under the advanced TPS control. Main simulation and experimental results are provided to validate the proposed fault diagnosis method.

Predicting traffic conditions for road segments is the prelude of working on intelligent transportation. Many existing methods can be used for short-term or long-term traffic prediction, but they focus more on regions than on road segments. The lack of fine-grained traffic predicting approach hinders the development of ITS. Therefore, MapLSTM, a spatio-temporal long short-term memory network preluded by map-matching, is proposed in this paper to predict fine-grained traffic conditions. MapLSTM first obtains the historical and real-time traffic conditions of road segments via map-matching. Then LSTM is used to predict the conditions of the corresponding road segments in the future. Breaking the single-index forecasting, MapLSTM can predict the vehicle speed, traffic volume, and the travel time in different directions of road segments simultaneously. Experiments confirmed MapLSTM can not only achieve prediction for road segments based a large scale of GPS trajectories effectively but also have higher predicting accuracy than GPR and ConvLSTM. Moreover, we demonstrate that MapLSTM can serve various applications in a lightweight way, such as cognizing driving preferences, learning navigation, and inferring traffic emissions.

In this paper, a novel multifunctional four-layer SPR structure (prism-metal- electro-optic crystal- external layer) is proposed which can not only measure the RI of external layer, but can also realize accurate measurement of the applied electric fields by fixing the external layer. When the structure is used for measuring the RI of external analytes, the metal and electro-optic crystal layers were optimized by FDTD solutions software. The thickness of Ag layer is 50 nm, and the optimum electro-optic crystal is BaTiO3 with a thickness of 15nm. The detection range is 1.33–1.37 and the sensitivity is up to 5602 nm/RIU. When the structure is used to measure electric field, simulation results shows an excellent linear relationship between the RI of the electro-optic crystal layer and the resonance wavelength, which is different from traditional SPR sensors. This special phenomenon greatly reduces the measurement error and improves measurement accuracy, for example a spectrometer with resolution of 1pm can achieve the minimum measurement of 7.729 × 10 - 2 V voltage.

With the increasing application of machine-to machine (M2M) communication through cellular networks, such as telematics, smart metering, point-of-sale terminals, and home security, more data traffice has been produced in the cellular network. Although many schemes have been proposed to reduce data traffic, they are inefficient in practical application due to poor adaption. In this paper, we focus on how to adaptively offload data traffic for cellular M2M networks. To this end, we propose an adaptive mobile data traffic offloading model (AOM). This model can decide whether to adopt opportunistic communications or communicate via cellular networks adaptively. In the AOM, we introduce traffic offloading rate (called TOR) and local resource consumption rate (called LRCR) and analyze them based on continue time Markov chain. Theory proof and extensive simulations demonstrate that our model is accurate and effective, and can adaptively offload data traffic of cellular M2M networks.

Moving objects gathering pattern represents a group events or incidents that involve congregation of moving objects, enabling the analysis of traffic system. However, effectively and efficiently discovering the specific gathering pattern turns to be a remaining challenging issue since the large number of moving objects will generate high volume of trajectory data. In order to address this issue, the authors propose a moving object gathering pattern retrieving method that aims to support the retrieving of gathering patterns based on spatio-temporal graph. In this method, firstly the authors use an improved R-tree based density clustering algorithm (RT-DBScan) to index the moving objects and collect clusters. Then, they maintain a spatio-temporal graph rather than storing the spatial coordinates to obtain the spatio-temporal changes in real time. Finally, a gathering retrieving algorithm is developed by searching the maximal complete graphs which meet the spatio-temporal constraints. To the best of their knowledge, effectiveness and efficiency of the proposed methods are outperformed other methods on both real and large trajectory data.

Continuum manipulators are deformable, passively compliant, and apt for manipulation in cluttered space. This paper addresses how to enable an $n$ -section continuum manipulator to probe an object, while gradually forming a whole-arm grasp in a cluttered environment. This approach is effective and efficient as evident from simulation and real experiments.

Moving objects gathering pattern represents a group events or incidents that involve congregation of moving objects, enabling the analysis of traffic system. However, how to improve the effectiveness and efficiency of the gathering pattern discovering method still remains as a challenging issue since the large number of moving objects will generate high volume of trajectory data. In order to address this issue, the authors propose a method to discovering the gathering pattern by analyzing the taxicab demand. This paper first introduces the concept of Taxicab Service Rate (TSR). In this method, they use the KS measures to test the distribution of TSR and calculate the mean value of the TSR of a certain time period. Then, the authors use a neural network based method Neural Network Gathering Discovering (NNGD) to detect the gathering pattern. The neural network is based on the knowledge of historical gathering pattern data. The authors have implemented their method with experiments based on real trajectory data. The results show the both effectiveness and efficiency of their method.

With the rapid growth of the mobile devices and the emergence of cloud computing, mobile cloud computing has gained widespread interest. In mobile cloud computing, a large-scale collection of mobile devices cooperate with each other to provide a cloud service at the edge. However, the improper mobile device selection has a negative effect on the quality of service. Existing methods are difficult to solve the problem, because they do not take the status and the historical characteristics of the mobile devices into consideration. This paper introduces a device status-aware and stability-aware mobile device selection method. Firstly, a model is designed to store the status and the historical characteristics of each mobile device. Secondly, an optimized cloud model is employed to evaluate the stability of each mobile device. Lastly, an optimal mobile device searching algorithm is presented to select the optimal mobile device. We provide an extensive evaluation of our method. The results show that our method can increase the quality of mobile cloud service compared with the traditional method.