Traffic Types Research Articles

Two-Stage Distillation-Aware Compressed Models for Traffic Classification

Traffic classification is indispensable for the Internet of Things (IoT) in intrusion detection and resource management. Deep learning (DL)-based strategies are the key tools for traffic classification due to high accuracy but still have some challenges. 1) It is hard to deploy complex DL models on resource-constraint IoT devices. 2) Performance is limited because of the ignorance of the similarity between IoT traffic. To address these issues, we propose lightweight but accurate models for traffic classification. First, we adopt a network-in-network basic model to reduce model size. Second, the basic model is trained with self-distilled response, feature map, and similarity among traffic types to enable its identification accuracy. Next, redundant filters are removed from the basic model to achieve compressed architectures. Then a teacher model updating scheme with knowledge distillation is proposed to train compressed models without compromising performance. Experimental results demonstrate that compared to the state-of-the-art deep packet model, the compressed model can achieve the highest accuracy, deal with imbalanced traffic, and reduce nearly 99% of computation overhead in two encrypted traffic classification scenarios, thus emphasizing its efficiency.

Application-Aware Scheduling for IEEE 802.15.4e Time-Slotted Channel Hopping Using Software-Defined Wireless Sensor Network Slicing.

Given the improvements to network flexibility and programmability, software-defined wireless sensor networks (SDWSNs) have been paired with IEEE 802.15.4e time-slotted channel hopping (TSCH) to increase network efficiency through slicing. Nonetheless, ensuring the quality of service (QoS) level in a scalable SDWSN remains a significant difficulty. To solve this issue, we introduce the application-aware (AA) scheduling approach, which isolates different traffic types and adapts to QoS requirements dynamically. To the best of our knowledge, this approach is the first to support network scalability using shared timeslots without the use of additional hardware while maintaining the application's QoS level. The AA approach is deeply evaluated compared with both the application traffic isolation (ATI) approach and the application's QoS requirements using the IT-SDN framework and by varying the number of nodes up to 225. The evaluation process took into account up to four applications with varying QoS requirements in terms of delivery rate and delay. In comparison with the ATI approach, the proposed approach enhanced the delivery rate by up to 28% and decreased the delay by up to 57%. Furthermore, even with four applications running concurrently, the AA approach proved capable of meeting a 92% delivery rate requirement for up to 225 nodes and a 900 ms delay requirement for up to 144 nodes.

Probability of rail break caused by out-of-round wheel loads

A simulation procedure to predict the probability of rail break due to a measured wheel load spectrum is presented. The load distribution includes a representative proportion of high-magnitude dynamic loads generated by out-of-round wheels. Linear elastic fracture mechanics is applied to determine the stress intensities of pre-existing rail head cracks in a continuously welded rail subjected to combined bending and temperature loading. Rail bending moments are evaluated using a validated time-domain model of dynamic vehicle–track interaction. The considered multi-dimensional stochastic parameter space includes field test data of dynamic loads from a wheel impact load detector and crack depths from eddy current data. Meta-models based on poly-harmonic splines are applied to reduce the computational cost of the analysis. Supported by the extensive field test data, the simulation procedure is demonstrated by investigating the influences of freight traffic type, track support stiffness and rail temperature on the probability of a rail break initiated at a pre-existing rail head crack.

HARE: Hybrid ARQ-Based Adaptive Retransmission Control Scheme for Synchronous Multi-Link in Wireless LANs

As the resolution of video becomes high and the number of mobile devices proliferates, the demand for video traffic is exploding. New traffic types, such as augmented reality (AR) and virtual reality (VR), which require high-throughput and low-latency, are also emerging. Two of the promising techniques discussed in increasing throughput and lowering latency are multi-link transmission and hybrid automatic repeat request (HARQ). However, in most cases of a device using a low-cost chipset, there are potential limitations to using HARQ on multi-link. This paper proposes a HARQ-based adaptive retransmission control scheme (HARE), which determines the optimal retransmission policy by casting it as a Markov decision process (MDP) for the devices undergoing the aforementioned limitations. Through this, we can decide whether to execute HARQ and select the adequate modulation and coding rate when a retransmission occurs. Through extensive analysis, we found that HARE can greatly increase throughput compared to the ARQ-only and HARQ-only retransmission methods with a small amount of buffer occupancy. This technique expects high-resolution video and brand-new services such as AR/VR to be enjoyed even by the devices with low-cost chipsets.

Comments on "The association between road traffic noise and type 2 diabetes: a systematic review and meta-analysis of cohort studies" by Wu, Shan et al., DOI https://doi.org/10.1007/s11356-023-25926-5.

Comments on "The association between road traffic noise and type 2 diabetes: a systematic review and meta-analysis of cohort studies" by Wu, Shan et al., DOI https://doi.org/10.1007/s11356-023-25926-5.

Resource-Constraint Network Selection for IoT Under the Unknown and Dynamic Heterogeneous Wireless Environment

We investigate the problem of network selection for the Internet of Things (IoT) to maximize the quality of experience (QoE) in a heterogeneous wireless environment. Different from the traditional network access approaches with the assumption that the network state information (NSI) is static and known a priori, a scenario where the NSI of networks is unknown and dynamic to IoT devices is considered. Due to hardware limitations in IoT, the device has a limited resource budget and consumes resources, e.g., the energy, during the process of network access. To maximize the cumulative QoE before resource exhausts, the device should learn and estimate the NSI of networks, and make appropriate decisions to balance the network selection and resource consumption. To address this issue, we formulate the problem as a combination of multi-armed bandit and optimization problems and propose two algorithms NCSA and NSA. Moreover, we consider the fact that the device has various traffic types and propose an algorithm MT-NSA. Theoretical analysis shows that the regret of all algorithms has a sub-linear relationship with the resource budget. The effectiveness is validated by simulations.

Modelling, simulation and performance evaluation of the IEEE 802.11e protocol with station mobility.

In this article, we present a parameterized Colored Petri Net (CPN) model of the IEEE 802.11e protocol for wireless communications with mobile stations. CPNs provide a graphical model for the modeling and analysis of concurrent systems, which can be parameterized by the use of constants, and thus they allow us to create more flexible models. Our CPN model captures the protocol's behavior, and the specific parameters used for the 802.11e protocol and the scenarios to be evaluated are captured by the CPN parameters. The model presented is flexible enough to cover full customization of traffic types, user mobility and collision avoidance protocols. In this model, there is an access point (AP) which is visible to all the stations, and we assume that due to physical restrictions, there are two range groups. All the stations in the same range group are visible to each other. The impact of mobility is then analyzed by studying a situation in which the stations move in a controlled way to the same range group. The simulation results demonstrate the impact on network performance for sensitive and insensitive traffic types, as well as the role of the RTS/CTS protocol in collision avoidance, especially when users are located in different regions. Specifically, we show how the performance improves in the different scenarios when the stations move to the same area, where they can see each other, and we also study the impact on the performance for each type of traffic.

Research on Traffic Sign Recognition Method Based on Generating Countermeasures Network

Nowadays, intelligent assisted driving technology is developing rapidly. In the field of intelligent assisted driving technology, traffic sign recognition is a crucial area which focuses on identifying input images and outputting corresponding traffic sign types in the images. However, due to the limitations of optical hardware in the image acquisition system or the poor environment in which the image acquisition system operates, the collected images may not be well recognized by classifiers used for traffic sign classification, despite containing traffic sign information. This could be due to low resolution of images, unclear images, and excessive noise. This article presents a modified generative adversarial network for reconstructing high-resolution traffic signs, specifically suited for the reconstruction of traffic sign images. The generated images and interpolated images are then sent into a convolutional neural network, which is a classifier, to complete the task of classifying traffic signs. The accuracy of the two sets of data is compared, and it is concluded that the classification accuracy of images reconstructed by the generative adversarial network is generally 15% higher than that of images processed by the interpolation method. This study not only proves that the low resolution of images containing traffic sign information can negatively affect subsequent traffic sign classification tasks, but also demonstrates that traffic sign recognition methods based on generating adversarial networks can indeed improve classification accuracy.

IARMTS: Design of an Interference-Aware Routing Model with Time Synchronization Capabilities for Dense Wireless Sensor Network Deployments

Performance of dense wireless sensor networks is often degraded due to communication interference and time synchronization issues. Existing machine learning & deep learning models that propose bioinspired & pre-emptive packet-analysis solutions for these tasks either have high complexity, or high deployment costs. Moreover, these models cannot be scaled for heterogeneous node & traffic types, which limits their applicability when applied to real-time scenarios. To overcome these issues, this text proposes design of an interference-aware routing model with time synchronization capabilities for dense wireless sensor network deployments. The network initially collects temporal clock states & packet delivery performance of different nodes on heterogeneous traffic scenarios. These traffic patterns are converted into frequency, entropy, Gabor, and Wavelet components. The converted components are used to train an ensemble set of Naïve Bayes (NB), k Nearest Neighbour (kNN), Multilayer Perceptron (MLP), and Support Vector Machine (SVM) classifiers. These classifiers assist in identification of optimal clock deviations and set of routing paths. These routing paths are further fine-tuned via use of a Bacterial Foraging Optimization (BFO) Model, which assists in identification of interference-aware paths. The BFO Model uses a temporal fitness function that fuses throughput, communication delay, energy levels, and packet delivery performance for different set of contextual communications. Due to which, the model is able to showcase lower end-to-end delay, higher throughput, lower energy consumption, and higher packet delivery performance when compared with existing routing methods under high density nodes & heterogeneous network scenarios. The model showcases 99% PDR, 18.3% lower delay, 19.5% higher energy efficiency and 10.4% lower delay levels when compared with existing methods.

Comparative Analysis of Machine Learning and Deep Learning Techniques for Intrusion Detection

Abstract: This comparative analysis examines the application of both machine learning and deep learning methods in network traffic classification. Network traffic classification holds significant importance in network security, traffic management, and Quality of Service provisioning. The analysis covers a range of popular machine learning techniques, such as Decision Tree, KNearest Neighbours, Naive Bayes, Logistic Regression, Multi-Layer Perceptron, and Feed Forward Neural Network with a sigmoid activation function. Each technique's strengths and weaknesses are discussed, along with the factors that influence the selection of a particular technique. Ultimately, the choice of machine learning approach depends on data characteristics, performance requirements, and available resources. The demand for prompt and accurate classification of Internet traffic has been steadily increasing, driven by the emergence of new applications in the field. Traditional approaches based on port numbers and packet payloads have become insufficient, prompting the adoption of pattern recognition techniques that leverage statistical flow-based features in training samples to classify unknown flows. To ensure real-time identification of traffic types, the chosen method must be capable of swift classification before the entire flow is completed. In this study, a supervised machine learning approach and deep learning techniques are proposed for the identification of various Internet applications. The proposed system exhibits the ability to detect application types based on just a few initial packets within each flow, enabling realtime operation. Promising results were achieved, with the Logistic Regression algorithm attaining the highest accuracy of 80.7%.

A comparative analysis of the semi-persistent and dynamic scheduling schemes in NR-V2X mode 2

Over the last years, the evolution of Vehicle-to-Everything (V2X) services from basic safety-related to enhanced V2X (eV2X) applications prompted the development of the 5G New Radio (NR)-V2X technology. Standardized by the Third Generation Partnership Project (3GPP) in Release 16, NR-V2X features a distributed resource allocation mode, known as Mode 2, that allows vehicles to autonomously select their transmission resources employing a Semi-Persistent Scheduling (SPS) or a Dynamic Scheduling (DS) scheme. The SPS approach relies on the periodic reservation of resources, whereas the DS scheme is a reservation-less solution that forces the selection of new transmission resources for every generated message. 3GPP standards do not indicate under which conditions each scheduling scheme should be used. In this context, this study analyzes and compares the performance of SPS and DS under different traffic types and Packet Delay Budget (PDB) requirements. Simulation results demonstrate that the SPS scheme represents the best solution for serving fixed size periodic traffic, whereas DS is more adequate for aperiodic traffic (of fixed or variable size). The study shows that the superiority of DS over SPS becomes more evident when tighter PDB requirements are considered, and that the performance of the DS scheme is independent of the PDB. It is also demonstrated that an adaptive scheduling strategy, which allows vehicles to select the scheduling scheme that best suits the type of generated traffic, is the best solution in mixed traffic scenarios where fixed size periodic traffic and variable size aperiodic traffic sources coexist.

Natural Factors Play a Dominant Role in the Short-Distance Transmission of Pine Wilt Disease

Pine wilt disease (PWD) is regarded as one of the most serious conifer diseases affecting pines worldwide. To date, an in-depth study of the driving mechanisms behind short-distance PWD spread is lacking. In this study, we collected PWD forest subcompartment data in Fushun, China, and analysed the effects of factors on the short-distance transmission of PWD; the analysed factors included the number of neighbouring PWD-infected forest subcompartments, the canopy density, the slope direction, and different traffic corridor types. The results suggested that the spatial spill-over effect of nearby PWD-infected subcompartments contributed the most to short-distance PWD transmission, with an impact of up to 78% on its propagation. The impact of the traffic corridor was 20%. With the help of a beetle vector, PWD can spread to nearby forest subcompartments, and this spatial PWD spill-over effect showed a linearly decaying trend as the distance to neighbouring subcompartments increased. Different traffic corridor types exhibited significant PWD transmission impact differences. County roadways and highways had great impacts, while others had relatively small impacts. For each additional 100 m of distance from a county roadway, highway, national, or provincial roadway, the PWD infection risks in forest subcompartments were reduced by 18%, 11%, 5%, and 3%, respectively. In this study, we quantified the influence of driving factors on the short-distance spread of PWD and provided a theoretical basis for the control of PWD transmission; the results obtained herein are critical for maintaining the ecological security of forests, promoting ecological forest management and stabilising forest carbon sinks.

Mediation of metabolic syndrome in the association between long-term co-exposure to road traffic noise, air pollution and incident type 2 diabetes

ObjectivesRecent studies have linked exposure to road traffic noise or air pollution with incident type 2 diabetes (T2D), but investigation on their co-exposure was limited and underlying mechanisms remain unclear. We hypothesized that long-term co-exposure to road traffic noise and air pollution increases the risk of incident T2D via the development of metabolic syndrome (MetS). MethodsThis prospective study included 390,834 participants in UK Biobank. Cumulative risk index (CRI), the health-based weighted levels of multiple exposures, was applied to characterize the co-exposure to 24-hour road traffic noise (Lden), particulate matter with aerodynamic diameter ≤ 2.5 µm (PM2.5), and nitrogen dioxide (NO2). Lden was modeled by the Common Noise Assessment Methods in Europe and air pollutant levels were measured by the Land Use Regression model at participants’ residential addresses. Incident T2D was ascertained through linkages to inpatient hospital records. MetS was defined by five (central obesity, triglycerides, HDL cholesterol, glucose, and blood pressure) or six factors (C-reactive protein additionally). Cox proportional hazard models were used to assess the association between environmental exposures and incident T2D, and mediation analyses were applied to investigate the role of MetS. ResultsAfter a median of 10.9 years of follow-up, 13,214 (3.4%) incident T2D cases were ascertained. The exposure to Lden, PM2.5, and NO2, as well as their co-exposure, were significantly associated with an elevated risk of incident T2D, with HRs of 1.03 (95%CI: 1.00, 1.05) per 3.5 dB(A) increase in Lden, 1.05 (95%CI: 1.01, 1.10) per 1.3 μg/m3 increase in PM2.5, 1.07 (95%CI: 1.02, 1.11) per 9.8 μg/m3 increase in NO2, and 1.06 (95%CI: 1.02, 1.09) per interquartile range increase in CRI. MetS significantly mediated 43.5%− 54.7% of the CRI-T2D relationship. ConclusionsLong-term co-exposure to road traffic noise and air pollution is associated with an elevated risk of incident T2D, which may partly be mediated by MetS.

Road Feature Detection for Advance Driver Assistance System Using Deep Learning.

Hundreds of people are injured or killed in road accidents. These accidents are caused by several intrinsic and extrinsic factors, including the attentiveness of the driver towards the road and its associated features. These features include approaching vehicles, pedestrians, and static fixtures, such as road lanes and traffic signs. If a driver is made aware of these features in a timely manner, a huge chunk of these accidents can be avoided. This study proposes a computer vision-based solution for detecting and recognizing traffic types and signs to help drivers pave the door for self-driving cars. A real-world roadside dataset was collected under varying lighting and road conditions, and individual frames were annotated. Two deep learning models, YOLOv7 and Faster RCNN, were trained on this custom-collected dataset to detect the aforementioned road features. The models produced mean Average Precision (mAP) scores of 87.20% and 75.64%, respectively, along with class accuracies of over 98.80%; all of these were state-of-the-art. The proposed model provides an excellent benchmark to build on to help improve traffic situations and enable future technological advances, such as Advance Driver Assistance System (ADAS) and self-driving cars.

Adaptive Random Access and Data Transmission Scheme With Mixed Traffic in NGSO Satellite Networks

In non-geostationary orbit satellite networks, the resource is limited and the traffic under each satellite is in variation due to the motion of satellites and the non-uniform traffic on the ground. Generally, the uplink communication is guaranteed by successful random access (RA) and sufficient data transmission (DT) resources for various traffic types. Different traffic types usually have different requirements for resources, and joint resource management for mixed traffic tries to meet the needs of various traffic types at the same time and may improve uplink resource efficiency. However, existing strategies have not given integrative consideration to jointly allocate resources for RA and DT under dynamic mixed traffic demands, which leads to inefficient network resource utilization. Considering such a problem, we propose an adaptive random access and data transmission scheme for NGSO satellite networks in the mixed traffic scenario, considering two types of users, i.e., real-time users and best-effort users. Specifically, the utility functions of different types of users are introduced, and a utility-based access control and resource allocation algorithm is developed by solving the network utility maximization problem. Our scheme aims to achieve optimal uplink network utilization by adaptively controlling random access and allocating resources according to the dynamic demands of mixed traffic. Simulation results show that the proposed scheme outperforms other methods in terms of network utility, and exhibits near-optimal performance.

Knacks of a hybrid anomaly detection model using deep auto-encoder driven gated recurrent unit

The cyber-attacks have recently posed a threat to national security; meanwhile, the pervasiveness of malware and cyber terrorism encumbers the beneficial utilization of the internet. Intrusion detection systems (IDS) can prevent such malevolent attacks. Inappropriate and redundant features affect the performance of IDS by slowing down the classification process and leading to incorrect decisions, specifically when dealing with big data. Therefore, in this study, we propose an auto-encoder and gated recurrent unit (GRU) based intrusion detection system (AE-GRU) to accurately, efficiently, and precisely classify network traffic. In the first step, the most relevant features are extracted from the auto-encoder to pass on to the GRU for traffic type classification. Classification of binary and multiclass have been carried out on the well-known NSL-KDD dataset. The AE-GRU is evaluated in terms of performance indices such as accuracy, precision, recall, F-score, MCC, DR, and FAR. The generalization of the proposed technique is also assessed on another dataset UNSW-NB15. Experimental results demonstrate that the AE-GRU outperforms existing methods in terms of all performance indices. Furthermore, the proposed model has also been statistically evaluated using a one-way ANOVA test. Results signify the potential utilization of the proposed technique in network traffic classification.

Vibration Influence on Structural Perspectives According to Pedestrian Traffic Type of Pedestrian Bridge

The vibration acceleration response and excellent frequency generated by pedestrian’s walking type of the steel composite wooden bridge were measured and analyzed. Based on the modified Reiher-Meister vibration curve, vibration influence evaluation on the structural perspectives of the pedestrian bridge was performed. The pedestrian’s walking types were divided into walking, running, rolling, and shaking, and eight experimental cases were set up by combining them with the number of users and moving direction. A total of 37 measurements were made. To evaluate the vibration serviceability of pedestrian bridges, the pedestrian’s walking type must be considered, and a standardized study on the pedestrian’s walking type will be necessary in the future.

Vibration Influence on the Pedestrians’ Perspectives According to Pedestrian Traffic Type of Pedestrian Bridge

The existing vibration serviceability evaluation of bridges focused on the structure. To evaluate vibration serviceability from a user perspective, vibration level felt by pedestrians was surveyed in a steel composite wooden bridge and analyzed to present the vibration serviceability evaluation method felt by the pedestrian. Such vibration was classified according to the walking type using the modified Reiher-Meister vibration curve, the existing evaluation method, and rolling up and down and rolling left and right in place were the most overestimated. As vibration is influenced by individuals, additional research is needed.

DeSlice: An Architecture for QoE-Aware and Isolated RAN Slicing.

Network slicing is considered a key feature of 5G and beyond cellular systems. It opens the door for new business models of mobile operators, enables new services, reduces costs with advanced infrastructure-sharing techniques, and improves heterogeneous traffic service. With slicing, the operators can tailor the network resources to the requirements of specific verticals, applications, and corresponding traffic types. To satisfy the heterogeneous quality of service (QoS) requirements of various slices, efficient virtualization and resource allocation algorithms are required. Such algorithms are especially crucial for the radio access network (RAN) because of the spectrum scarcity. This article develops DeSlice, a novel architecture for RAN slicing. DeSlice enables efficient real-time slicing algorithms that satisfy heterogeneous QoS requirements of the slices and improve the quality of experience for their end users. The article illustrates the advantages of DeSlice by considering the problem of the joint service of cloud VR, video, and web traffic. It develops the algorithms using DeSlice architecture and application-to-network communication. With simulations, it shows that, together, the architecture and the algorithms allow greatly improving the QoE for these traffics significantly.

Using Bidirectional Encoder Representations from Transformers (BERT) to classify traffic crash severity types

Traffic crashes are a critical safety concern. Many studies have attempted to improve traffic safety by performing a wide range of studies on safety topics with the application of diverse statistical and machine learning models. The data elements contained in police-reported crash narrative information are not routinely analyzed with coded and structured crash data. In the recent years, unstructured textual contents in traffic crash narratives have been investigated by many researchers. However, most of these studies are basic text mining applications and often the dataset is limited in size. This study applied an advanced language model Bidirectional Encoder Representations from Transformers (BERT) to classify traffic injury types by using a dataset of over 750,000 unique crash narrative reports. The models have an 84.2% ±0.5 predictive accuracy and an Area Under the receiver operating Curve (AUC) of 0.93 ± 0.06 per class. Overall, the findings can assist safety engineers and analysts in determining the causes of a crash. The classification of crash injury types using a language model like BERT is a valuable tool for identifying additional factors that contribute to crashes, which can identify new areas for safety countermeasures and support the development of new safety strategies.