Traffic Types Research Articles

Comparative Evaluation of VAEs, VAE-GANs and AAEs for Anomaly Detection in Network Intrusion Data

With cyberattacks growing in frequency and sophistication, effective anomaly detection is critical for securing networks and systems. This study provides a comparative evaluation of deep generative models for detecting anomalies in network intrusion data. The key objective is to determine the most accurate model architecture. Variational autoencoders (VAEs), VAE-GANs, and adversarial autoencoders (AAEs) are tested on the NSL-KDD dataset containing normal traffic and different attack types. Results show that AAEs significantly outperform VAEs and VAE-GANs, achieving AUC scores up to 0.96 and F1 scores of 0.76 on novel attacks. The adversarial regularization of AAEs enables superior generalization capabilities compared to standard VAEs. VAE-GANs exhibit better accuracy than VAEs, demonstrating the benefits of adversarial training. However, VAE-GANs have higher computational requirements. The findings provide strong evidence that AAEs are the most effective deep anomaly detection technique for intrusion detection systems. This study delivers novel insights into optimizing deep learning architectures for cyber defense. The comparative evaluation methodology and results will aid researchers and practitioners in selecting appropriate models for operational network security.

TSN Network Scheduling—Challenges and Approaches

Time-Sensitive Networking (TSN) is a set of Ethernet standards aimed to improve determinism in packet delivery for converged networks. The main goal is to provide mechanisms that enable low and predictable transmission latency and high availability for demanding applications such as real-time audio/video streaming, automotive, and industrial control. To provide the required guarantees, TSN integrates different traffic shaping mechanisms including 802.1Qbv, 802.1Qch, and 802.1Qcr, allowing for the coexistence of different traffic classes with different priorities on the same network. Achieving the required quality of service (QoS) level needs proper selection and configuration of shaping mechanisms, which is difficult due to the diversity in the requirements of the coexisting streams under the presence of potential end-system-induced jitter. This paper discusses the suitability of the TSN traffic shaping mechanisms for the different traffic types, analyzes the TSN network configuration problem, i.e., finds the optimal path and shaper configurations for all TSN elements in the network to provide the required QoS, discusses the goals, constraints, and challenges of time-aware scheduling, and elaborates on the evaluation criteria of both the network-wide schedules and the scheduling algorithms that derive the configurations to present a common ground for comparison between the different approaches. Finally, we analyze the evolution of the scheduling task, identify shortcomings, and suggest future research directions.

Evaluation of the QoS policy model of an ordinary 5G smart city cluster with predominant URLLC and eMBB traffic

A typical element of the smart city’s information and communication space is a 5G cluster, which is focused on serving both new and handover requests because it is an open system. In an ordinary 5G smart city cluster, Ultra-Reliable Low-Latency Communications (URLLC) and enhanced Mobile BroadBand (eMBB) traffic types prevail. The formation of an effective QoS policy for such an object (taking into account the potentially active slicing technology) is an urgent problem. As a baseline, this research considers a Quality of Service (QoS) policy with constraints for context-defined URLLC and eMBB classes of incoming requests. Evaluating the QoS policy instance defined within the framework of the basic concept requires the formalization of both a complete qualitative metric and a computationally efficient mathematical apparatus for its calculation. The article presents accurate and approximate methods of calculating such quality parameters as the probability of loss of typed requests and the utilization ratio of the communication resource, which depend on the implementation of the estimated QoS policy. At the same time, the original parametric space includes both fixed characteristics (amount of available communication resources, load according to request classes) and controlled characteristics due to the specifics of the implementation of the basic QoS concept. The paper empirically proves the adequacy of the presented mathematical apparatus for evaluating the QoS policy defined within the scope of the research. Also, in the proposed qualitative metric, a comparison of the author’s concept with a parametrically close analogue (the well-known QoS policy scheme, which takes into account the phenomenon of reservation of communication resources), determined taking into account the reservation of communication resources, was made. The results of the comparison testify in favour of the superiority of the author’s approach in the proposed metrics.

P4SQA: A P4 Switch-Based QoS Assurance Mechanism for SDN

Data center networks include a wide variety of service categories, which leads to a complex network topology. As a result, the QoS of different service flows cannot be guaranteed. In this paper, we propose a QoS assurance mechanism based on P4 switches (P4SQA) to make full use of network resources. P4SQA includes two functions: in-network traffic classification and in-network queue management. To reduce the communication overhead and computational pressure on the Software Defined Network (SDN) controller, we offload part of the neural computation from the control plane to the data plane. A fully connected neural network-based traffic classification method is deployed in the programmable data plane to classify different traffic types. In addition, the optimized active queue management algorithm CoDeL is integrated into the P4 switch (PCoDeL) to prevent network congestion while guaranteeing QoS more effectively. The evaluation results show that traffic classification accuracy has improved, especially in terms of the F1_score, which reached 0.97, significantly better than other state-of-the-art methods. Moreover, compared to traditional queue management algorithms, PCoDeL can improve network throughput and reduce transmission delay, which verifies the effectiveness and superiority of P4SQA in ensuring QoS.

Modelling, Simulation, and Performance Analysis of Intra-Vehicular Heterogeneous Networks

Vehicles today use a variety of network segments operated by different technologies and protocols within the car (CAN, LIN, Automotive Ethernet, MOST, FlexRay, etc.) to exchange data between different control modules, sensors, and actuators. The exchange of information between other network domains (heterogeneous networks) is enabled through various interconnection points called gateways/bridges. The resulting performance depends on its interconnection structure, network segment traffic aggregation scheme, and medium access technique. Although protocols such as CAN, LIN, FlexRay, and Ethernet have been used in network design for some time, performance modeling and analysis are still needed given the variety of traffic types and sources, new application limitations, and especially the lack of formal verification of network performance for different network scenarios and configurations. This paper presents an end-to-end throughput and delay performance analysis for a reference intra-vehicular network scenario. These models have been validated through simulations in which high correlation values were obtained from 98.7400 to 99.9999, with a low mean square error. The validation cases show that for different LIN, CAN, and Ethernet network configurations, the performance threshold values defined for most current vehicle applications are preserved. However, if the network configuration is modified, the proposed analytical models can be used to formally verify the corresponding performance and delay changes and thus validate whether or not the application requirements are met.

Effect of rail traffic noises on the perception of the acoustic environment in office spaces

This study conducted a laboratory experiment to investigate the effect of rail traffic noise on the perception of the acoustic environment in office spaces. The study considered two types of experiment stimuli, single noise, and combined noise. The single noise stimuli consisted of silence (SL), air conditioning noise (AC), irrelevant speech (SP), and six types of traffic noise (Tr). Traffic noise included road (R), maglev (Ma), tram (T), conventional train (C), high-speed train (H), and metro (Me). The combined noise stimuli used air conditioning sound and speech as background noise and combined with six types of traffic noise to form the experimental stimuli. Thirty subjects took part in laboratory experiments in which the acoustic environment was rated on two aspects: noise disturbance and acoustic comfort. Under single noises, the results showed that the noise source had a significant negative effect on the evaluation of the acoustic environment. The influence of traffic noises varied with the traffic type, which was lower and higher than that of irrelevant speech and air conditioning noise, respectively. Under combined acoustic conditions, acoustic comfort was affected by the main effects of background noise type, traffic noise type, and traffic noise level. In addition to the main effects, noise disturbance was also influenced by the interaction between traffic noise level and background noise type, and between traffic noise level and traffic noise type. Keywords: rail traffic noise; office space; acoustic environment; acoustic comfort; noise disturbance

Masking effect of water sound on road and rail traffic noises in office soundscape

This study represents a laboratory experiment focusing on the effect of water sound on masking traffic noises in an office soundscape. Subjective evaluations and physiological responses of thirty-two participants exposed to the combination of water sound and traffic noises were measured. Silence and pure traffic noises were used as the reference conditions. Significant masking effects of water sound were observed on both the subjective evaluation of the sound environment and the physiological responses to traffic noises. To improve the subjective evaluation of the sound environment, the water sound was found to have stronger masking effect on rail traffic noises than the road traffic noise. To reduce the noise annoyance, SNRs at -3dB and 3dB was found to be the most effective for the conventional train noise and the high-speed train noise, respectively. Water sound at low sound level (SNR=-3dB) was also found to be effective in improving the acoustic comfort for the conventional train noise. Effect of water sound in masking the physiological impact of traffic sounds was found to be limited which depended on the traffic noise type and participants characteristics. Comparing to EDA, HR were found to be more easily affected by the water sound.

Reconfigurable optical crosshaul architecture for 6G radio access networks

The radio access network (RAN) architecture is undergoing a significant evolution to support the next-generation mobile networks and their emerging applications. To realize scalable and sustainable deployment and operations, RAN needs to consider the requirements of 6G and beyond wireless technologies such as ultra densification of cells, higher data rates, ubiquitous coverage, and new radio spectrum in the millimeter-wave band. This calls for a careful redesign of every aspect of RAN, including its crosshaul. The crosshaul is an important network segment in future RAN, capable of transporting diverse traffic types with varying stringent requirements within RAN. The crosshaul towards 6G is envisioned to be highly intelligent, reconfigurable, and adaptable to dynamic service requirements and network conditions. To this end, we propose a software defined network (SDN)-enabled reconfigurable optical crosshaul architecture (ROCA) that supports heterogeneous crosshaul transport technologies and dynamic functional splittings. ROCA enables efficient and intelligent control of the crosshaul data plane. The proposed architecture with a set of the next-generation RAN (NG-RAN) transport interfaces is evaluated using network models built on the ns-3 network simulator. Simulation results demonstrate the strengths and weaknesses of different crosshaul interfaces in agreement with the understanding of respective NG-RAN interfaces from the literature, which validates the modeling accuracy. We then demonstrate the reconfigurability of the architecture using a dynamic scenario with different reconfiguration strategies for meeting the user and network demands. The results indicate that ROCA serves as a scalable and flexible foundation for supporting high-capacity delay-stringent RAN that can be used in 6G and beyond wireless technologies.

Network slice allocation for 5G V2X networks: A case study from framework to implementation and performance assessment

Empowered by the capabilities provided by fifth generation (5G) mobile communication systems, vehicle-to-everything (V2X) communication is heading from concept to reality. Given the nature of high-mobility and high-density for vehicle transportation, how to satisfy the stringent and divergent requirements for V2X communications such as ultra-low latency and ultra-high reliable connectivity appears as an unprecedented challenging task for network operators. As an enabler to tackle this problem, network slicing provides a power tool for supporting V2X communications over 5G networks. In this paper, we propose a network resource allocation framework which deals with slice allocation considering the coexistence of V2X communications with multiple other types of services. The framework is implemented in Python and we evaluate the performance of our framework based on real-life network deployment datasets from a 5G operator. Through extensive simulations, we explore the benefits brought by network slicing in terms of achieved data rates for V2X, blocking probability, and handover ratio through different combinations of traffic types. We also reveal the importance of proper resource splitting for slicing among V2X and other types of services when network traffic load in an area of interest and quality of service of end users are taken into account.

Deep Learning-Based Efficient Analysis for Encrypted Traffic

To safeguard user privacy, critical Internet traffic is often transmitted using encryption. While encryption is crucial for protecting sensitive information, it poses challenges for traffic identification and poses hidden dangers to network security. As a result, the precise classification of encrypted network traffic has become a crucial problem in network security. In light of this, our paper proposes an encrypted traffic identification method based on the C-LSTM model for encrypted traffic recognition by leveraging the power of deep learning. This method can effectively extract spatial and temporal features from encrypted traffic, enabling accurate identification of traffic types. Through rigorous testing and evaluation, our system has achieved an impressive accuracy rate of 96.4% on the widely used ISCXVPN2016 dataset. This achievement demonstrates the effectiveness and reliability of our method in accurately classifying encrypted network traffic. By addressing the challenges posed by encrypted traffic identification, our research contributes to enhancing network security and privacy protection.

Malicious traffic detection in multi-environment networks using novel S-DATE and PSO-D-SEM approaches

The rapid advancement of network architectures, protocols, and tools poses significant challenges to network security, especially due to the use of AI-based tools by cybercriminals. It is crucial to develop a versatile malicious traffic detection system capable of identifying attacks across diverse traffic types. This paper presents an enhanced system for detecting malicious traffic in multi-environment (M-En) networks, including IoT, SDN, and traditional IP-based traffic. Existing techniques often under-utilize diversity in network traffic, limiting their effectiveness. To address this, we propose a comprehensive approach that combines the power of Synthetic Data Augmentation TEchnique (S-DATE) and Particle Swarm Optimizer (PSO)-based Diverse-Self Ensemble Model (D-SEM). Our approach proposes the M-En dataset, a combination of three different network architectures datasets including InSDN, UNSWNB-15, and IoTID-20 that accurately represent real-world scenarios. S-DATE is then employed to address imbalanced data distribution in novel generated M-En dataset, enabling better model convergence and enhancing the detection rate of normal and abnormal traffic. Additionally, we introduce PSO-D-SEM, a novel ensemble model that leverages the diversity provided by PSO to handle the complexity of M-En networks. The PSO-D-SEM combines individual models trained on a subset of the M-En dataset, resulting in improved overall performance. The experimental results demonstrate the superiority of our enhanced system, achieving a significant accuracy score of 0.989. Further, we also deploy a statistical T-test to demonstrate the significance of the proposed PSO-D-SEM approach in comparison with state-of-the-art methods.

A Method for Analyzing Navigation Flows of Health Website Users Seeking Complex Health Information with Google Analytics

People are increasingly seeking complex health information online. However, how they access this information and how influential it is on their health choices remains poorly understood. Google Analytics (GA) is a widely used web analytics tool and it has been used in academic research to study health information-seeking behaviors. Nevertheless, it is rarely used to study the navigation flows of health websites. To demonstrate the usefulness of GA data, we adopted both top-down and bottom-up approaches to study how web visitors navigate within a website delivering complex health information about stem cell research using GA’s device, traffic and path data. Custom Treemap and Sankey visualizations were used to illustrate the navigation flows extracted from these data in a more understandable manner. Our methodology reveals that different device and traffic types expose dissimilar search approaches. Through the visualizations, popular web pages and content categories frequently browsed together can be identified. Information on a website that is often overlooked but needed by many users can also be discovered. Our proposed method can identify content requiring improvements, enhance usability and guide a design for better addressing the needs of different audiences. This paper has implications for how web designers can use GA to help them determine users’ priorities and behaviors when navigating complex information. It highlights that even where there is complex health information, users may still want more direct and easy-to-understand navigations to retrieve such information.

Urban vertical air pollution gradient and dynamics investigated with low-cost sensors and large-eddy simulations

A network of five low-cost air pollution sensor (LCS) nodes was deployed vertically on the exterior of the H. C. Ørsted Institute at the University of Copenhagen, Denmark, to investigate the transport of pollution from the road below. All LCS nodes measured PM2.5, NO2, and O3 at 1-min time resolution, and one of them also measured noise. Traffic was monitored with a webcam, where traffic type and levels were derived using a machine-learning algorithm. We investigated how well traffic-related air pollution, noise, and real-time traffic counts serve as proxies for one another. The correlations between NO2, noise, and traffic count exhibited relatively low values when considering all the data. However, these correlations significantly increased under southwesterly wind direction and low wind speed, reaching R2 = 0.40 for NO2 and noise, R2 = 0.51 for NO2 and traffic volume, and R2 = 0.70 for noise and traffic volume. These results indicate a common source, namely traffic, for all three parameters. The five LCS nodes spanning 25 m vertically had extremely low intervariability with minimum R2-values of 0.98 for PM2.5, 0.89 for NO2, and 0.97 for O3. The system could not detect a vertical gradient in pollution levels. Large-eddy simulation model runs using the PALM model system generally supported the lack of gradient observed in measured observations. Under slightly unstable stratification, concentration remained relatively constant with height for southwesterly and southerly winds. Conversely, winds from the north, west, and northwest showed an increase in concentration with height. For other wind directions, the concentration decreased with height by approximately 40 % to 50 %, which is not as strong as for neutral stratification, attributed to enhanced vertical mixing under unstable stratification. Based on the measurements and modeling, we conclude that the vertical concentration profile is very sensitive to stratification, and under these conditions, the concentration outside the window of a fifth-floor office is almost the same as for an office on the ground floor.

Detection and utilization of new-type encrypted network traffic in distributed scenarios

Encrypted Network Traffic Classification (ENTC) is a crucial task in network management. The existing ENTC schemes usually imply two hypotheses. First, adopting a centralized processing mode. Second, the number of network traffic categories is fixed. However, it is usually unreasonable to collect all network traffic to the centralized node for processing which may cause network congestion. In addition, new traffic types emerge in endlessly, and the existing models cannot classify them. This paper studies the ENTC problem in distributed scenarios, where multiple monitoring nodes coexist, and both the existing types and new types of encrypted traffic exist at the same time. These nodes cooperate to train models, which can alleviate the challenges of limited numbers and types faced by each single node. In order to solve the proposed ENTC problem, firstly, we propose a feature extraction method in the distributed scenario. Then a detection method for new type traffic and a classification method for existing type traffic are proposed. Finally, a globally consistent automatic category labeling method and classification model updating method for new types encrypted traffic are proposed. We also take Security Information and Event Management (SIEM) as an example to discuss the managerial implications.

An RSU-Assisted Hybrid Emergency Message Broadcasting Protocol for VANETs

In vehicular ad hoc networks (VANETs), emergency messages broadcasting has been considered as one of the important parts for safety related applications. The efficiency and reliability of the emergency message broadcasting are severely affected by the network structure, message redundancy, channel contention, etc. To address this problem, in this paper, we propose an RSUs-assisted hybrid emergency message broadcasting (RA-HEMB) protocol for two-way grid roads in urban VANETs. In doing so, the broadcasting target region is first generated based on the instantaneous traffic status and types of the emergency messages. To balance the deliver latency and the reliability, an adaptive forwarding nodes selection scheme based on position is proposed to dynamically determine the forwarding area according to different vehicle densities and vehicle-to-vehicle (V2V) communication ranges. Then, based on the forwarding nodes selection scheme and roadside unit (RSU) deployment, an RA-HEMB protocol is established by properly utilizing both the wired links via RSUs and the wireless links to broadcast the emergency message in hybrid vehicular networks. Simulation results show that the proposed RA-HEMB can greatly shorten the broadcasting time needed to notice the nodes within the target region performance compared to the benchmark protocols.

Predicting Relationship Between Traffic Accidents with Related Injuries

Abstract: Traffic collisions are one of the world's 14 most crucial issues right now since they cause countless fatalities, severe injuries, and financial losses every year. The difficulty of creating accurate forecasting models, Traffic accident severity is important for transportation networks. This study creates models to choose a number of important features and build a model for classifying injury severity. These models are produced utilizing various Machine Learning techniques. The data on traffic accidents, trained and methods for unsupervised machine learning are also used. The primary aim is to draw a connection between the many kinds of injuries and traffic accident types. The study's conclusions imply that unsupervised learning methods may be effective in predicting the severity of harm from traffic accidents.

Modeling of Traffic Information and Services for the Traffic Control Center in Autonomous Vehicle-Mixed Traffic Situations

Achieving fully autonomous driving requires seamless collaboration between advanced autonomous driving and road infrastructure technologies. As the proportion of autonomous vehicles (AVs) increases, challenges may arise from their insufficient knowledge of the behavior of traffic objects and inability to effectively drive short distances. Therefore, traffic control centers that can proactively control these issues in real time are essential. In this study, first, the terminology is defined and the types of AV-mixed Traffic Information that a traffic control center needs to efficiently collect, store, and analyze to accommodate the coexistence of AVs and conventional vehicles are identified. Second, a generic notation for an AV-mixed Traffic Information model is defined and the results of modeling each AV-mixed Traffic Information type are presented. Third, an AV-mixed Traffic Information services model that included the names, operations, input/output messages, and relationships of all services is suggested. Finally, the importance of the service functionalities is evaluated through a survey. This study will serve as an initial guideline for the design, construction, and operation of traffic control centers and will help proactively address issues that may arise from the interaction between AVs and conventional vehicles on the road. Moreover, it contributes to identifying the types of traffic information and services that traffic control centers must provide in the era of AV-mixed traffic and suggests future directions for analysis and utilization of traffic information.

Life cycle assessment of road network infrastructure maintenance phase while considering traffic operation and environmental impact

The road maintenance has great influence on traffic operation and environment in many aspects from the perspective of life cycle assessment (LCA). To clarify this problem, LCA of road network infrastructure maintenance phase while considering traffic operation and environmental impact is researched in this paper. Firstly, the evaluation objective, quantification method and evaluation content for LCA are determined. Then, a list of factors related to road network infrastructure maintenance cost is categorized as fixed and variable costs. The relationship between evaluation results and maintenance scale is established. Quantitative models are established for the evaluation of maintenance management and traffic operations in the maintenance phase of road network infrastructure. Besides, the impact of maintenance activity on maintenance management and traffic operations is explored. The effects of different maintenance durations and traffic control types on traffic condition are investigated. It is found that there is a significant scale economy effect in maintenance costs, which provides a calculation basis for optimizing maintenance strategies. This research improves the analysis method for the LCA of road network infrastructure maintenance phase while considering traffic operation and environmental impact, which provides an important reference and guidance for the efficient and low-carbon operation of road network infrastructure.

Effects of Traffic Clutter and Road Type on Road Hazard Detection

While operating a manually driven or even a highly automated vehicle, drivers may need to quickly perceive and respond to road hazards. In this study, we investigate several factors that might influence drivers’ ability to rapidly recognize road hazards in real-world settings. Using the publicly available Road Hazard Stimuli dashcam footage set, we divided stimuli into videos from highway and non-highway environments. We also developed the concept of “traffic clutter,” which is the area of the video screen containing other vehicles, and divided videos into those with high or low traffic clutter. Half the video clips showed road hazards occurring, and the other half showed another snippet from the same dashcam footage without a road hazard. Forty participants viewed these short, 233 ms video clips and responded whether they detected a road hazard or not. Using signal detection analyses, we show that drivers behave differently when responding to road hazards, depending on the context. First, participants were well above chance at detecting road hazards despite the short presentation times. Further, participants had higher sensitivity (d’) to road hazards when viewing videos with high traffic clutter. In terms of participants’ response criterion (c), we found that in both non-highway and high traffic clutter scenarios, drivers were more willing to commit false alarm rather than miss hazards. This difference in hazard response strategies may reflect biases drivers have developed during their time on the road.

Defining The Decisive Factors on Purchase and Comparing Feature Importance Methods

Online retail companies focus on two activities for getting revenue in their businesses and to survive in the market. First activity is increasing traffic of the online shopping platform and second activity is converting this traffic to revenue for the company. Marketing facilities try to attract customers to the online shopping platforms at great costs. Because of the costs of getting traffic, it is crucial to make customers order. Online shopping platforms need to understand which factors are decisive on customer purchase decision. In this study, which factors are decisive on consumer purchase decisions will be studied on an e-commerce retail platform from Turkey, Hepsiburada. Which of these factors are most decisive try to be defined: traffic source type (google, campaign, direct etc.) of the customer, customer persona or segment, which types of page or page components has been seen, product position on the page, does the customer benefited from campaign or discount, product review scores and counts, has the product recommended or not. In this study, data will be gathered from Hepsiburada transactions stored in google's big query environment. Performance problems will be solved via SQL optimization and other methods. Data quality issues will be fixed to get consistent results. Then statistical methods, supervised machine learning and deep learning methods will be applied to data for getting feature importances. Importance value of the features will show which factor decisive on customer purchase decision. Feature importance values will be compared and evaluated according to method, model results. Hyperparameter tunings is applied to the methods. Also, the model performances will be compared and evaluated. This study uses and compares 7 methods and there is no comprehensive study in literature in terms of method variety.