Search

Traffic prediction plays an essential role in intelligent transportation system. Accurate traffic prediction can assist route planing, guide vehicle dispatching, and mitigate traffic congestion. This problem is challenging due to the complicated and dynamic spatio-temporal dependencies between different regions in the road network. Recently, a significant amount of research efforts have been devoted to this area, especially deep learning method, greatly advancing traffic prediction abilities. The purpose of this paper is to provide a comprehensive survey on deep learning-based approaches in traffic prediction from multiple perspectives. Specifically, we first summarize the existing traffic prediction methods, and give a taxonomy. Second, we list the state-of-the-art approaches in different traffic prediction applications. Third, we comprehensively collect and organize widely used public datasets in the existing literature to facilitate other researchers. Furthermore, we give an evaluation and analysis by conducting extensive experiments to compare the performance of different methods on a real-world public dataset. Finally, we discuss open challenges in this field., to be published in IEEE Transactions on Intelligent Transportation Systems

Accurate and timely traffic flow prediction is crucial for intelligent transportation systems (ITS). Recent advances in graph-based neural networks have achieved promising prediction results. However, some challenges remain, especially regarding graph construction and the time complexity of models. In this paper, we propose a multi-stream feature fusion approach to extract and integrate rich features from traffic data and leverage a data-driven adjacent matrix instead of the distance-based matrix to construct graphs. We calculate the Spearman rank correlation coefficient between monitor stations to obtain the initial adjacent matrix and fine-tune it while training. As to the model, we construct a multi-stream feature fusion block (MFFB) module, which includes a three-channel network and the soft-attention mechanism. The three-channel networks are graph convolutional neural network (GCN), gated recurrent unit (GRU) and fully connected neural network (FNN), which are used to extract spatial, temporal and other features, respectively. The soft-attention mechanism is utilized to integrate the obtained features. The MFFB modules are stacked, and a fully connected layer and a convolutional layer are used to make predictions. We conduct experiments on two real-world traffic prediction tasks and verify that our proposed approach outperforms the state-of-the-art methods within an acceptable time complexity.

Communication networks grow exponentially in this globalization era; thus, the network traffic modelling and prediction plays a crucial role in network management and security warning. Solely, the multi-step network traffic prediction may involve greater errors hence worsening prediction performance. To overcome this problem, an optimized echo state network model with selective error compensation is proposed. In the optimized echo state network-based multi-step prediction model, an improved fruit–fly optimization algorithm based on cloud model (named LVCMFOA) is used to select optimum values of four key parameters of the model. The proposed LVCMFOA algorithm uses the levy-flight function to redefine the generation of the fruit–fly population, which can randomly change the search radius and help getting out of a possible local optimal solution and prevent local optimum. To reduce the calculation time but improve the prediction accuracy simultaneously, a sophisticated selective error compensation strategy employing the variable sliding window technology is proposed so as to avoid the error accumulation problem in the multi-step prediction. The effectiveness of the proposed method is verified by applying it to Henon mapping chaotic series, Mackey–Glass chaotic series and two public network traffic data sets all known in the literature.

Auto machine learning recently has been introduced as a trending technique for learning applications, including smart transportation. In this study, we focus on applying auto-machine learning for hyperparameter tuning to learn traffic datasets at the main regions of highway systems. Particularly, deep learning models have been recently introduced as emergent methods for traffic prediction. However, training deep learning models requires expensive works (e.g., time-consuming and human expertise), especially in terms of determining the configurations of hyperparameters in the models. In this regard, this paper introduces an automated framework for hyperparameter tuning to learn traffic datasets at an ecosystem in terms of reducing time-consuming tasks. Specifically, we first propose the HyperNet framework, using advanced data science techniques (e.g., Bayesian optimization and meta-learning) for the automated hyperparameter search process. Then, a deep learning model with the long short term memory network based on the HyperNet framework has presented for learning the temporal variation of traffic datasets at main regions of highway traffic systems. Regarding the experiment, we take data from the Korean highway system into account as a case study to evaluate the proposed approach. The evaluation indicates promising results of the proposed framework for learning multiple datasets of the traffic highway systems.

Traffic flow mismanagement is a significant challenge in all countries especially in crowded cities. An alternative solution is to utilize smart technologies to predict traffic flow. In this study, frequency spectrum describing traffic sound characteristics is used as an indicator to predict the next five-minute vehicle density. Sound frequency and vehicle intensity are collected during a thirteen-hour data gathering. The collected sound intensity and frequency are then used to learn three machine-learning models - support vector machine, artificial neural network, and random forest and to predict vehicle intensity. It was found out that the performances of the three models based on root-mean-square-error values are 12.97, 16.01, and 10.67, respectively. These initial and satisfactory results pave a new way to predict traffic flow based on traffic sound characteristics which may serve as a better alternative to conventional features.

International audience; Route planning represents a major challenge with a substantial impact on safety, economy, and even climate. An ever-growing urban population caused a significant increase in commuting times, therefore, stressing the prominence of efficient real-time route planning. In essence, the goal is to compute the fastest route to reach the target location in a realistic environment where traffic conditions are time-evolving. Consequently, a large volume of traffic data is potentially required and the route continuously updated. We thereby address the rerouting problem to answer questions such as when, how often, and where is rerouting worthwhile. We base our study on a real dataset, comprising the travel times of the road segments of New York, London, and Chicago, collected over three months. By exploiting this dataset, we implement an optimal algorithm, able to mimic ideal predictions of road segment speeds in the network. Thereby, allowing us to compute the lower bound of travel-time to serve as a reference against other routing techniques. Mainly, we quantify the achieved travel-time gain of a static, no rerouting , and continuous rerouting strategies. Surprisingly, we find that traffic conditions are sufficiently stable for short time windows, and rerouting a vehicle is very seldom useful when exploiting accurate statistics at departure time. Typically, real-time rerouting should only be triggered during rush hours, for long routes, passing through well-identified road segments.

Internet of Things (IoT) is an expanded application of Internet, which are used to provide various services for users. Up to now, IoTs have received more attention, because it can provide ubiquitous connectivity for various devices. With the development of IoTs, its security has become a main issue. Attackers use various techniques to implement cyber attacks for the IoT, which threats the users’ privacy seriously. As a security mechanism, intrusion detection techniques can detect various illicit behaviors before attackers invade the network. An intrusion detection system can implement effective defense functions to keep the network away from attacks. This paper proposes an intrusion detection algorithm based on deep reinforcement learning, which pursued the trends of traffic flows by extracting statistical features of prior network traffic for traffic prediction at first. Then, we use traffic predictors to employ intrusion detection. The evaluations verify the effectiveness of our algorithm in detecting Distributed Denial of Service attacks (DDoS).

Traffic prediction on road networks is highly challenging due to the complexity of traffic systems and is a crucial task in successful intelligent traffic system applications. Existing approaches mostly capture the static spatial dependency relying on the prior knowledge of the graph structure. However, the spatial dependency can be dynamic, and sometimes the physical structure may not reflect the genuine relationship between roads. To better capture the complex spatial‐temporal dependencies and forecast traffic conditions on road networks, a multi‐step prediction model named Spatial‐Temporal Attention Wavenet (STAWnet) is proposed. Temporal convolution is applied to handle long time sequences, and the dynamic spatial dependencies between different nodes can be captured using the self‐attention network. Different from existing models, STAWnet does not need prior knowledge of the graph by developing a self‐learned node embedding. These components are integrated into an end‐to‐end framework. The experimental results on three public traffic prediction datasets (METR‐LA, PEMS‐BAY, and PEMS07) demonstrate effectiveness. In particular, in the 1 h ahead prediction, STAWnet outperforms state‐of‐the‐art methods with no prior knowledge of the network.

In this article, we propose a new protection scheme for backbone networks to guarantee high service availability. The presented scheme does not require any reconfiguration immediately after the failure (i.e., it is proactive). At the same time, it does not require any reserved backup network resources either. To achieve these seemingly contradictory goals, we utilize the recent advancements in Machine Learning (ML) to implement a network intelligence that periodically re-allocates the unused capacity as protection bandwidth to meet the service availability requirements of each connection. Our goal is achieved by two components (1) predicting the traffic for the next period on each link, and (2) intelligently selecting the best fit dedicated protection scheme for the next period depending on the estimated unused (spare) bandwidth and the previous service availability violations. Note that re-allocating protection bandwidth affects neither the operational connections nor the current best practice of operators to over-provision network bandwidth to support elephant flows. Finally, we provide a case study on the real traffic from Energy Sciences Network (ESnet), a high-speed, international scientific backbone network. The key benefit of our framework is that adaptively utilizing the over-provisioned bandwidth for spare capacity is sufficient to improve the availability from three-nines to five-nines (in ESnet for the 30 examined connections). The drawback is negligible bandwidth limitations; the user perceives a minor and very temporal bandwidth limitation in less than 0.1% of the time.

Modeling network traffic is an endeavor actively carried on since early digital communications, supporting a number of practical applications, that range from network planning and provisioning to security. Accordingly, many theoretical and empirical approaches have been proposed in this long-standing research, most notably, Machine Learning (ML) ones. Indeed, recent interest from network equipment vendors is sparking around the evaluation of solid information-theoretical modeling approaches complementary to ML ones, especially applied to new network traffic profiles stemming from the massive diffusion of mobile apps. To cater to these needs, we analyze mobile-app traffic available in the public dataset MIRAGE-2019 adopting two related modeling approaches based on the well-known methodological toolset of Markov models (namely, Markov Chains and Hidden Markov Models ). We propose a novel heuristic to reconstruct application-layer messages in the common case of encrypted traffic. We discuss and experimentally evaluate the suitability of the provided modeling approaches for different tasks: characterization of network traffic (at different granularities, such as application, application category, and application version), and prediction of network traffic at both packet and message level. We also compare the results with several ML approaches, showing performance comparable to a state-of-the-art ML predictor (Random Forest Regressor). Also, with this work we provide a viable and theoretically sound traffic-analysis toolset to help improving ML evaluation (and possibly its design), and a sensible and interpretable baseline.

Traffic prediction is a core problem in the intelligent transportation system and has broad applications in the transportation management and planning, and the main challenge of this field is how to efficiently explore the spatial and temporal information of traffic data. Recently, various deep learning methods, such as convolution neural network (CNN), have shown promising performance in traffic prediction. However, it samples traffic data in regular grids as the input of CNN, thus it destroys the spatial structure of the road network. In this paper, we introduce a graph network and propose an optimized graph convolution recurrent neural network for traffic prediction, in which the spatial information of the road network is represented as a graph. Additionally, distinguishing with most current methods using a simple and empirical spatial graph, the proposed method learns an optimized graph through a data-driven way in the training phase, which reveals the latent relationship among the road segments from the traffic data. Lastly, the proposed method is evaluated on three real-world case studies, and the experimental results show that the proposed method outperforms state-of-the-art traffic prediction methods.

The timely access to accurate traffic data is essential to the development of intelligent traffic systems. However, the existing traffic prediction methods cannot achieve satisfactory results, mainly because of three factors: the structure is too simple to extract deep features; many external factors are overlooks, such as weather and traffic incidents; the nonlinearity of traffic flow is not well handled. To solve the problem, this paper improves the deep belief network (DBN), a deep learning method, for accurate traffic prediction under poor weather. Firstly, the data of poor weather and traffic data were collected from IoV, rather than induction coils in traditional methods. Next, the support vector regression (SVR) was introduced to improve the classic DBN. In the improved DBN, the underlying structure is a traditional DBN that learns the key features of traffic data in an unsupervised manner, and the top layer is an SVR that performs supervised traffic prediction. To verify its effectiveness, the improved DBN was applied to predict the traffic data based on the traffic data from the control center of an expressway and the weather data from local monitoring stations, in comparison with the autoregressive integrated moving average (ARIMA) model and the traditional neural network. The experimental results show that the improved DBN controlled the traffic prediction error within 9%, and maintained good robustness despite the extension of the time interval. To sum up, this paper provides an effective way to predict traffic flow under poor weather, shedding new light on the application of deep learning in traffic prediction.

Efficient machine learning algorithms capturing both spatial and temporal dependencies offer promising results for many fields including traffic prediction. However, their performance is still limi...

Traffic prediction helps mitigate the impact of traffic congestion. The accuracy of traffic predictions depends on the availability of the data used for the prediction as well as the prediction model. Data from fixed traffic detectors is only available at certain locations. On the other hand, connected vehicles can provide Floating Car Data (FCD) at any location and time. However, FCD may not be available at all vehicles, and this can impact predictions since the FCD may not reflect the state of all traffic. This impact is larger when predicting traffic density or flow, and existing studies generally use FCD to predict traffic speed or travel time only. This study proposes a traffic prediction model that can accurately predict the three fundamental traffic variables (traffic density, flow, and speed) using FCD and an error recurrent convolutional neural network that takes as input the three variables. These are estimated using FCD and data from induction loops. These estimates depend on the penetration rate of FCD, so we propose a method to locally and dynamically estimate this penetration rate. This method improves the estimation of the traffic variables, and hence their prediction. The proposed model is used to analyze the impact of the FCD penetration rate on the prediction of the traffic variables. We show how our proposal reduces the amount of FCD needed to improve the prediction obtained with data from traffic detectors. We show that our proposal only requires FCD from 4% of the vehicles to improve the prediction accuracy achieved with traffic detectors. Augmenting this percentage increases the accuracy of our model for the three traffic variables. However, we also show that our prediction model reduces the FCD sample size (or FCD penetration rate) needed to achieve prediction accuracy levels close to that obtained if all vehicles provided FCD.

It is challenging to deal with the Internet congestion problem because of several factors such as ever-growing traffic and distributed network architecture. The congestion problem can be solved or alleviated by various methods, including rate control, bandwidth-guarantee routing and bandwidth reservation. We use the term broad-sensed Internet congestion control and avoidance (BICC&A) to generally denote all of the above methods. Most BICC&A solutions depend on or benefit from the knowledge of network conditions, including traffic status (type and volume), available bandwidth and topology. In this paper, we present a comprehensive survey of the applications of machine learning to network condition acquirement methods for BICC&A and specific BICC&A methods. First, we provide an overview of the background knowledge of BICC&A and machine learning. Then, we provide detailed reviews on the applications of machine learning techniques to network condition acquirement methods for BICC&A and to specific BICC&A methods. Finally, we outline important research opportunities.

Road traffic forecasting is crucial in Intelligent Transportation Systems (ITS). To achieve accurate results, it is necessary to model the dynamic nature and the complex non-linear dependencies governing traffic. The goal is particularly challenging when the prediction involves more than just one traffic variable. This paper proposes a novel multi-task learning model, called AST-MTL, to perform multi-horizon predictions of the traffic flow and speed at the road network scale. The strategy combines a multilayer fully-connected neural network (FNN) and a multi-head attention mechanism to learn related tasks while improving generalization performance. The model also includes the graph convolutional network (GCNs) and the gated recurrent unit network (GRUs) to extract the spatial and temporal features of traffic conditions. Our experiments employ new sets of GPS data, called OBU data, to perform traffic prediction in the freeway and urban contexts. The experimental results prove our model can effectively perform multi-horizon traffic forecasting for different types of roads and outperform state-of-the-art models., SCOPUS: ar.j, info:eu-repo/semantics/published

Traffic prediction is necessary for management departments to dispatch vehicles and for drivers to avoid congested roads. Many traffic forecasting methods based on deep learning have been proposed in recent years, and their main aim is to solve the problem of spatial dependencies and temporal dynamics. This paper proposes a useful dynamic model to predict the urban traffic volume by combining fully bidirectional LSTM, a complex attention mechanism, and external features, including weather conditions and events. First, we adopt bidirectional LSTM to obtain temporal dependencies of traffic volume dynamically in each layer, which is different from the hybrid methods combining bidirectional and unidirectional approaches. Second, we use a more elaborate attention mechanism to learn short-term and long-term periodic temporal dependencies. Finally, we collect weather condition and event information as external features to further improve the prediction precision. The experimental results show that the proposed model improves the prediction precision by approximately 3-7 percent on the NYC-Taxi and NYC-Bike datasets compared to the most recently developed method and is therefore a useful tool for urban traffic prediction.

Traffic speed forecasting plays an increasingly essential role in successful intelligent transportation systems. However, this still remains a challenging task when the accuracy requirement is dema...

Knowledge about future optical network traffic can be beneficial for network operators in terms of decreasing an operational cost due to efficient resource management. Machine Learning (ML) algorithms can be employed for forecasting traffic with high accuracy. In this paper we describe a methodology for predicting traffic in a dynamic optical network with service function chains (SFC). We assume that SFC is based on the Network Function Virtualization (NFV) paradigm. Moreover, other type of traffic, i.e. regular traffic, can also occur in the network. As a proof of effectiveness of our methodology we present and discuss numerical results of experiments run on three benchmark networks. We examine six ML classifiers. Our research shows that it is possible to predict a future traffic in an optical network, where SFC can be distinguished. However, there is no one universal classifier that can be used for each network. Choice of an ML algorithm should be done based on a network traffic characteristics analysis.

Road transportation is the largest and complex nonlinear entity of the traffic management system. Accurate prediction of traffic-related information is necessary for an effective functioning of Intelligent Transportation System (ITS). It is still a challenge for the departments of transportation to choose an appropriate prediction technique for the ITS applications. That is, a user must be able to utilize the disseminated information effectively by the forecasting models. This paper provides a detailed survey of the latest forecasting technologies and contributes to understand the key concept behind the prediction approaches. To provide guidelines to the decision-maker, this paper reviews multifaceted techniques developed by various authors for traffic prediction. We start classifying each technique into four categories namely, Machine Learning (ML), Computational Intelligence (CI), Deep Learning (DL), and hybrid algorithms. Many have conducted survey using model-driven or data-driven methods. We are the first to explore the area of traffic prediction based on the advances in multifaceted techniques proposing algorithmic approaches for key traffic characteristics in the forecasting process. The role of dependent factors in the prediction are analyzed thoroughly. We have analyzed each algorithm chronologically based on various traffic traits. The approaches are summarized based on the rational usage and performance of each technique. The analysis led to several research queries, and the appropriate responses are provided based on our detail survey. Finally, it is confirmed that currently, CI-MLs and DL hybrid techniques outperforms the rest in the field of traffic prediction. Ultimately suggested open challenges and future direction to explore the capability of DL and hybrid techniques further in the field of traffic prediction.

The objective of the study is to predict traffic flow under unusual conditions by using a deep learning model. Conventionally, machine-learning-based traffic prediction is frequently carried out. Model learning requires large amounts of training data; however, collecting sufficient samples is a challenge in the event of traffic incidents. To address this challenge, large amounts of traffic data were generated by performing traffic simulations under various traffic incidents. These data were used as training data, and a deep learning model with graph convolution and input of traffic incident information features was proposed. Subsequently, the prediction accuracy was compared with other models such as long short-term memory, which is typically used in traffic prediction. The results demonstrated the superiority of the proposed model in representing phenomena with strong spatio-temporal dependencies, such as traffic flow, and its effectiveness in traffic prediction.

Traffic control has been an issue for a long time from the past. The modern world demands Technology. Now a days cars are one of the main methods of improvement in technology. Intelligent Traffic System is also known as Intelligent Transportation System apply communication and information technology to find the solution for the Traffic control issues. Intelligent Transportation System represents the main problem in transportation. ITS is a program .it is used to improve the efficiency of transportation through advanced technologies by using sensors and communication. Some of the problems like Traffic congestion, Low safety can be solved through this Intelligent Transportation System by Using the Latest techniques in traffic management.ITS is improved by using wireless and wireline communication-based information, control and electronic technologies. Now a days overspeeding is a key issue in the traffic control system to overrule the issue. Dophler Phenomenon is used for speed measurement.

The next-generation cellular systems, including fifth-generation cellular systems (5G), are empowered with the recent advances in artificial intelligence (AI) and other recent paradigms. The internet of things (IoT) and the tactile internet are paradigms that can be empowered with AI solutions and integrated with 5G systems to deliver novel services that impact the future. Machine learning technologies (ML) can understand examples of nonlinearity from the environment and are suitable for network traffic prediction. Network traffic prediction is one of the most active research areas that integrates AI with information networks. Traffic prediction is an integral approach to ensure security, reliability, and quality of service (QoS) requirements. Nowadays, it can be used in various applications, such as network monitoring, resource management, congestion control, network bandwidth allocation, network intrusion detection, etc. This paper performs time series prediction for IoT and tactile internet delays, using the k-step-ahead prediction approach with nonlinear autoregressive with external input (NARX)-enabled recurrent neural network (RNN). The ML was trained with four different training functions: Bayesian regularization backpropagation (Trainbr), Levenberg–Marquardt backpropagation (Trainlm), conjugate gradient backpropagation with Fletcher–Reeves updates (Traincgf), and the resilient backpropagation algorithm (Trainrp). The accuracy of the predicted delay was measured using three functions based on ML: mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE).

With the increase of traffic prediction models, there has become an urgent need to develop a standardized framework to implement and evaluate these methods. This paper presents LibCity, a unified, comprehensive, and extensible library for traffic prediction, which provides researchers with a credible experimental tool and a convenient development framework. In this library, we reproduce 42 traffic prediction models and collect 29 spatial-temporal datasets, which allows researchers to conduct comprehensive experiments in a convenient way. To accelerate the development of new models, we design unified model interfaces based on unified data formats, which effectively encapsulate the details of the implementation. To verify the effectiveness of our implementations, we also report the reproducibility comparison results of LibCity, and set up a performance leaderboard for the four kinds of traffic prediction tasks. Our library will contribute to the standardization and reproducibility in the field of traffic prediction. The open source link of LibCity is https://github.com/LibCity/Bigscity-LibCity.

Nowadays, with the rapid development of IoT (Internet of Things) and CPS (Cyber-Physical Systems) technologies, big spatiotemporal data are being generated from mobile phones, car navigation systems, and traffic sensors. By leveraging state-of-the-art deep learning technologies on such data, urban traffic prediction has drawn a lot of attention in AI and Intelligent Transportation System community. The problem can be uniformly modeled with a 3D tensor (T, N, C), where T denotes the total time steps, N denotes the size of the spatial domain (i.e., mesh-grids or graph-nodes), and C denotes the channels of information. According to the specific modeling strategy, the state-of-the-art deep learning models can be divided into three categories: grid-based, graph-based, and multivariate time-series models. In this study, we first synthetically review the deep traffic models as well as the widely used datasets, then build a standard benchmark to comprehensively evaluate their performances with the same settings and metrics. Our study named DL-Traff is implemented with two most popular deep learning frameworks, i.e., TensorFlow and PyTorch, which is already publicly available as two GitHub repositories https://github.com/deepkashiwa20/DL-Traff-Grid and https://github.com/deepkashiwa20/DL-Traff-Graph. With DL-Traff, we hope to deliver a useful resource to researchers who are interested in spatiotemporal data analysis., This paper has been accepted by CIKM 2021 Resource Track

The evolution of IoT has come with the challenge of connecting not only a massive number of devices, but also providing an always wider variety of services. In the next few years, a big increase in the number of connected devices is expected, together with an important increase in the amount of traffic generated. Never before have wireless communications permeated so deeply in all industries and economic sectors. Therefore, it is crucial to correctly forecast the spectrum needs, which bands should be used for which services, and the economic potential of its utilization. This paper proposes a methodology for spectrum forecasting consisting of two phases: a market study and a spectrum forecasting model. The market study determines the main drivers of the IoT industry for any country: services, technologies, frequency bands, and the number of devices that will require IoT connectivity. The forecasting model takes the market study as the input and calculates the spectrum demand in 5 steps: Defining scenarios for spectrum contention, calculating the offered traffic load, calculating a capacity for some QoS requirements, finding the spectrum required, and adjusting according to key spectral efficiency determinants. This methodology is applied for Colombia’s IoT spectrum forecast. We provide a complete step-by-step implementation in fourteen independent spectrum contention scenarios, calculating offered traffic, required capacity, and spectrum for cellular licensed bands and non-cellular unlicensed bands in a 10-year period. Detailed results are presented specifying coverage area requirements per economic sector, frequency band, and service. The need for higher teledensity and higher spectral efficiency turns out to be a determining factor for spectrum savings.

With the complexity of the network system rapidly increasing, network traffic prediction has great significance for the safety pre-warning of the network load, network management and control, and improvement of the quality of the network service. In this paper, the time series analysis is used for the network traffic prediction, and a prediction method combined with an optimized unscented Kalman filter (UKF) by an improved fruit fly algorithm (IFOA) and echo state network (ESN) is proposed, which is named by IFOA-UKF-ESN. The researches mainly solve the problem that the prediction accuracy might be greatly affected by the actual network traffic data with unknown and time-varying noises. UKF is used to train the best state vector (formed by spectral radius, scale of the reservoir, scale of the input units and connectivity rate) of ESN; and the proposed IFOA algorithm is proposed to optimize the weights of the predicted state value and the covariance in UKF, which makes UKF have adaptive ability for unknown and time-varying noise. Three actual network traffic data sets with different Gaussian white noise distributions are constructed for experiments, and the experimental results show that the proposed prediction method makes an average improvement by reducing at least 20.60%, 43.23% and 41.85% of RMSE, at least 23.66%, 52.38% and 47.50% of MAE, and at least 23.58%, 52.10% and 47.28% of MAPE, which verify the effectiveness of the proposed method. National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61773108] The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors are grateful for financial support from the National Natural Science Foundation of China (Grant No.61773108).

Congestion often hinders human mobility. This problem occurs due to the constant increase in vehicles every year. Reliable predictions of traffic conditions would allow drivers to choose their routes to avoid traffic jams while providing the police with traffic management strategies. Therefore, this paper tests the ability of various machine learning methods to predict traffic conditions. The study assesses Neural Networks, Bayes Classifier, Decision Trees, SVM, Deep Neural Network, and Deep Learning. Of these methods, the Decision Tree, Deep Neural Network, and Bayes Classifier show the highest performance in predicting traffic conditions using static data testing. However, in dynamic testing to assess the growth of knowledge, the performance of the Knowledge-Growing Decision Tree tends to decrease as the training data grows. Its performance decreased 3.89 points (88.24% to 84.35%) in accuracy, and 7.55 points (76.25% to 68.70%) for each precision, recall, and F1 Score. Conversely, the Knowledge-Growing Deep Neural Network and Bayes Classifier had a better performance than Decision Tree. The performances of Knowledge-Growing Deep Neural Network increased slightly by 0.35 points (93.38% to 93.73%) for accuracy and 0.69 points (86.77% to 87.64%) in other measurements. Although its performance increased, the processing time takes very long, namely 139452.76 seconds and 318832.80 seconds for sub-scheme (a) and (b), respectively. Meanwhile, the Knowledge-Growing Bayes Classifier offers a greater performance increase of 2.3 points (80.52% to 82.82%) for the accuracy and 4.6 points (65.63% to 61.03%) for the other performance measurements. In addition, it also scored better for processing time, as predictions only take 3 seconds using sub-scheme (a), and 7 seconds when using sub-scheme (b). Therefore, the paper proposes the Knowledge-Growing Bayes Classifier to predict rapidly changing traffic conditions. This method outperform the others. These can be attributed to its ability to 1) adjust to ever-changing the traffic conditions; 2) predict the result as soon as the data are acquired; and 3) make decentralized predictions.

Eco-approach and departure (EAD) systems in connected vehicles are capable of providing the speed recommendations to drivers so that vehicles can pass through successive intersections at the appropriate instants to save time and energy. However, most of existing control strategies and global optimization algorithms for EAD systems have extremely high requirements for computational resources, which make EAD systems not easy to popularize and apply to real vehicles. To overcome this problem, this paper designs a traffic flow prediction model based on deep learning regression machine, and establishes a dynamic effective red-light duration model based on traffic flow queuing effect. To facilitate real-time update of the optimal speed, a constrained optimization model is proposed as an approximation approach, which can obtain similar optimal results to that of the pseudo-spectral method while greatly reducing calculating time. The effectiveness of the proposed control algorithm for EAD system on passing intersections and energy saving has been validated on the real vehicle under real-world traffic environment. Compared with the uninformed driver, the proposed EAD system saves 2% time and 4.6% energy on average in case that driver is informed and 8% time and 12.1% energy in case of autonomous driving.

The 21st century is a high-tech information era in which our lives are closely linked by computer networks. Hence, how to effectively supervise networks and reduce the frequency of network security incidents has now become a research hotspot in cyberspace. Specifically, researchers have shown an increased interest in predicting the network traffic before any untoward incident happens. Optimization and decomposition technologies are the core components of network traffic prediction model which plays an important role in network management. This article discusses past network traffic prediction research and critically examines the optimization and decomposition technologies used in the model, lists the model parameter structure based on the research methodology, the data set used, the evaluation criteria and so on. By comparison, digging out the Particle Swarm Optimization (PSO) algorithm and the Variational Mode Decomposition (VMD) decomposition technique will effectively solve the network traffic model predictive difficulties that have proven to be crucial to improving predictive accuracy and convergence speed strategy.The comprehensive review reveals that PSO and VMD are the most suitable optimization algorithm and decomposition technology for network traffic prediction modeling.

Anomaly detection for large scale cellular networks can be used by network operators to optimize network performance and enhance mobile user experience. This paper aims at detecting user anomalies from spatio-temporal cell phone activity data. We design an approach combining time series analysis and machine learning to extract the traffic patterns of areal units. This approach can cluster areal units with similar traffic patterns and segment a city into distinct groups. Then, in grouped-areas, we use a clustering technique to detect anomalous behaviors of the cellular network and verify the accuracy of the results using ground truth information collected from online sources. The results indicate that anomalies are associated with abruptly high or unexpected traffic demand at a specific location and time. In addition, we obtain anomaly-free data by removing anomalous data and train a decomposed traffic prediction model. It is observed that the prediction model trained with anomaly-free data can achieve lower normalized mean square error (NMSE), i.e., higher prediction accuracy, than the model trained with anomalous data.

Wireless cellular traffic prediction is a critical issue for researchers and practitioners in the 5G/B5G field. However, it is very challenging since the wireless cellular traffic usually show high nonlinearities and complex patterns. Most existing wireless cellular traffic prediction methods, lacking abilities of modeling the dynamic spatial-temporal correlations of wireless cellular traffic data, thus cannot yield satisfactory prediction results. To improve the accuracy of 5G/B5G cellular network traffic prediction, more cross-domain data was considered, a cross-service and regional fusion transfer learning strategy (Fusion-transfer) based on the spatial-temporal cross-domain neural network model (STC-N) was proposed. Multiple cross-domain datasets were integrated. The training accuracy of the target service domain based on the data characteristics of its source service domain according to the similarity between services and the similarity between different regions was improved, so the predictive performance of the model was enhanced. The experimental results show that the prediction accuracy of the traffic prediction model is significantly improved after the integration of multiple cross-domain datasets, the RMSE performance of SMS, Call and Internet service can be improved about 8.39%, 13.76% and 5.7% respectively. In addition, compared with the existing transfer strategy, the RMSE of the three services can be improved about 2.48%~13.19%.

Deep learning-based techniques are the state of the art in road traffic prediction or forecasting. Several deep neural networks have been proposed to predict the traffic but they have not been evaluated under common datasets. Current studies analyze their capacity to predict road traffic in general but do not focus on their capacity to predict the formation of congestions. This is critical for avoiding congestions or mitigate their negative impact. This paper progresses the current state of the art by presenting a comprehensive comparison of the state-of-the-art deep neural networks for road traffic prediction. The comparison is conducted using the same real traffic datasets, and under normal and congested traffic conditions. The evaluation includes new deep neural networks and error recurrent models. Our study first demonstrates that accurately predicting the traffic overall does not imply that a deep neural network can accurately predict the traffic when congestions are being formed. This reinforces the idea that prediction techniques must also be evaluated under congestion conditions. Our analysis also shows that exploiting the spatiotemporal evolution of the traffic (and not just the temporal one) provides better prediction accuracy overall and in particular under congestion conditions. The study also demonstrates that error recurrent models outperform deep neural networks that do not utilize an error feedback both under normal and congested traffic conditions. In particular, our study shows that the error recurrent model eRCNN is the deep learning technique that achieves to date the best traffic prediction accuracy. It is also important emphasizing that error recurrent models achieve better prediction accuracy with shallower neural networks and therefore lower computational cost.

Mobile edge computing (MEC) becomes popular as it offers cloud services and functionalities to the edge devices, to enhance the quality of service (QoS) of end-users by offloading their computationally intensive tasks. At the same time, the rise in the number of internet of things (IoT) objectives poses considerable cybersecurity issues owing to the latest rise in the existence of attacks. Presently, the development of deep learning and hardware technologies offers a way to detect the present traffic condition, data offloading, and cyber-attacks in edge networks. The utilization of DL models finds helpful in several domains in which the MEC provides the decisive beneficiary of the approach for traffic prediction and attack detection since a large quantity of data generated by IoT devices enables deep models to learn better than shallow approaches. In this view, this paper presents a new DL based traffic prediction with a data offloading mechanism with cyber-attack detection (DLTPDO-CD) technique. The proposed model involves three major processes traffic prediction, data offloading, and attack detection. Initially, bidirectional long short term memory (BiLSTM) based traffic prediction to enable the proficient data offloading process. Then, the adaptive sampling cross entropy (ASCE) technique is executed to maximize the network throughput by making decisions related to offloading users to the WiFi system. Finally, a deep belief network (DBN) optimized by a barnacles mating optimizer (BMO) algorithm called BMO-DBN is applied as a detection tool for cyberattacks in MEC. Extensive simulation is carried out to ensure the proficient performance of the DLTPDO-CD model. The experimental outcome stated the superiority of the presented model over the compared methods under different dimensions.

This chapter summarizes the background and motivation for why people need to make every effort to predict network traffic and user mobility. To address all the technical aspects related to traffic and mobility predictions with the emphasis on the machine learning based solutions, the chapter first presents the problem formation and a brief overview of existing prediction methods in terms of modelling, characterization, complexity, and performance. It presents some deep learning based schemes for traffic and mobility prediction. The chapter also introduces random connectivity long short‐term memory – a model that reduces computational cost by randomly removing some neural connections – and its performance in traffic prediction. Then it shows three long short‐term memory‐based user‐mobility prediction schemes, two of which take into account the spatial dependence on the user's movement trajectory, thus combining convolutional neural networks to expect to achieve better prediction performance.

The use of base station (BS) sleep modes is one of the most studied approaches for the reduction of the energy consumption of radio access networks (RANs). Many papers have shown that the potential energy saving of sleep modes is huge, provided the future behavior of the RAN traffic load is known. This paper investigates the effectiveness of sleep modes combined with machine learning (ML) approaches for traffic forecast. A portion of a RAN is considered, comprising one macro BS and a few small cell BSs. Each BS is powered by a photovoltaic (PV) panel, equipped with energy storage units, and a connection to the power grid. The PV panel and battery provide green energy, while the power grid provides brown energy. Our study examines the impacts of different prediction models on the consumed energy mix and on QoS. Numerical results show that the considered ML algorithms succeed in achieving effective trade-offs between energy consumption and QoS. Results also show that energy savings strongly depend on traffic patterns that are typical of the considered area. This implies that a widespread implementation of these energy saving strategies without the support of ML would require a careful tuning that cannot be performed autonomously and that needs continuous updates to follow traffic pattern variations. On the contrary, ML approaches provide a versatile framework for the implementation of the desired trade-off that naturally adapts the network operation to the traffic characteristics typical of each area and to its evolution. pub

Traffic prediction is of great importance to traffic management and public safety, and very challenging as it is affected by many complex factors, such as spatial dependency of complicated road networks and temporal dynamics, and many more. The factors make traffic prediction a challenging task due to the uncertainty and complexity of traffic states. In the literature, many research works have applied deep learning methods on traffic prediction problems combining convolutional neural networks (CNNs) with recurrent neural networks (RNNs), which CNNs are utilized for spatial dependency and RNNs for temporal dynamics. However, such combinations cannot capture the connectivity and globality of traffic networks. In this paper, we first propose to adopt residual recurrent graph neural networks (Res-RGNN) that can capture graph-based spatial dependencies and temporal dynamics jointly. Due to gradient vanishing, RNNs are hard to capture periodic temporal correlations. Hence, we further propose a novel hop scheme into Res-RGNN to utilize the periodic temporal dependencies. Based on Res-RGNN and hop Res-RGNN, we finally propose a novel end-to-end multiple Res-RGNNs framework, referred to as “MRes-RGNN”, for traffic prediction. Experimental results on two traffic datasets have demonstrated that the proposed MRes-RGNN outperforms state-of-the-art methods significantly.

Traffic prediction is a major component of any traffic management system. With the increase in data sources and advancement in connectivity, data analysis and machine learning approaches for traffic prediction have gained a lot of attention. Most of the existing data analysis approaches in traffic prediction rely on aggregated inputs such as flow and density, with limited studies using the individual vehicle-level data. The time-space diagram of the vehicles can be constructed from the connected vehicles’ data. This plot is comprehensive and contains all the information about traffic flow dynamics at both microscopic and macroscopic levels. Accordingly, this study introduces a deep learning-based methodology to directly predict the traffic state based on the time-space diagram with the use of convolutional neural networks (CNN). The time-space diagram is directly used as the input to the traffic prediction model using a CNN. The prediction capability of the proposed model is compared with multilayer perceptron, support vector regression, and autoregressive integrated moving average, and the results indicate a superior capability of CNN in predicting flow and density across all possible values of these parameters.

With the rapid increase of vehicle ownership in China, toll collection stations on expressways have become some of the most congested areas. Compared with a manual toll collection system (M...

Rapid growth of network traffic causes the need for the development of new network technologies. Artificial intelligence provides suitable tools to improve currently used network optimization methods. In this paper, we propose a procedure for network traffic prediction. Based on optical networks’ (and other network technologies) characteristics, we focus on the prediction of fixed bitrate levels called traffic levels. We develop and evaluate two approaches based on different supervised machine learning (ML) methods—classification and regression. We examine four different ML models with various selected features. The tested datasets are based on real traffic patterns provided by the Seattle Internet Exchange Point (SIX). Obtained results are analyzed using a new quality metric, which allows researchers to find the best forecasting algorithm in terms of network resources usage and operational costs. Our research shows that regression provides better results than classification in case of all analyzed datasets. Additionally, the final choice of the most appropriate ML algorithm and model should depend on the network operator expectations.

In this paper, we formulate the problem of traf-fic prediction in optical networks. We then design a machine learning approach based on Graph Convolutional Network and the Generative Adversarial Network to enable efficient network states forecasting. The main focus is on detecting the peak traffic in networks that can affect the routing decisions. We validate our results using pseudorealistic datasets generated in a custom simulator and real networks provided by the network operator. The findings confirm our approach’s efficiency for optimizing both the real-time routing and long-term network design decisions.

This article first proposes an adaptive traffic scheduling strategy for optoelectronic hybrid data centers. The strategy is composed of a deep bidirectional LSTM-based traffic prediction model and a prediction-assisted traffic scheduling method. The simulation results confirm that the presented method can achieve non-congested intra-data center traffic scheduling and higher network performance even under heavy traffic conditions.