Artificial intelligence-based vehicular traffic flow prediction methods for supporting intelligent transportation systems

Boukerche, Azzedine; Tao, Yanjie; Sun, Peng

doi:10.1016/j.comnet.2020.107484

Cited by 120 publications

(34 citation statements)

References 88 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A recomendação ITU-T Y.3172 prevê a introdução de mecanismos de aprendizado de máquina para o gerenciamento e a orquestração funcionalidades nas próximas gerações de rede 5 . Nesse sentido, modelos estatísticos baseados em séries temporais são os métodos clássicos com bom desempenho sempre escolhidos para realizar predições [Boukerche et al, 2020], sobretudo em fluxos de rede, por possuírem uma boa capacidade analítica e uma implementação com baixo custo computacional. A previsão de séries é um campo essencial do aprendizado de máquina aplicado a redes 5G [Chakraborty et al, 2020].…”

Section: As Ferramentas Baseadas Em Modelos Estatísticosunclassified

Segurança em Redes 5G: Oportunidades e Desafios em Detecção de Anomalias e Predição de Tráfego Baseadas em Aprendizado de Máquina

Barbosa¹,

Bezerra²,

Medeiros³

et al. 2021

Minicursos Do XXI Simpósio Brasileiro De Segurança Da Informação E De Sistemas Computacionais

View full text Add to dashboard Cite

This chapter focuses on approaching and contextualizing the security of the fifthgeneration (5G) mobile networks, discussing network anomaly detection techniques through hybrid tools. Classical techniques for prediction, such as time series regression analysis and the Hidden Markov Model, are revisited. New anomaly detection and traffic prediction techniques based on deep learning are presented, such as recurrent neural networks, neural networks with long short-term memory, and convolutional neural networks. Finally, the challenges and new paradigms of the next-generation networks (6G) are presented. We also present a case study with a practical exercise to develop an example of anomaly detection and traffic prediction through open source and free tools. ResumoEste capítulo aborda e contextualiza a segurança das redes móveis de quinta geração (5G), discutindo técnicas de predição de tráfego e detecção de anomalias em redes através de ferramentas híbridas. Revisitam-se técnicas clássicas para predição de tráfego, como análise de regressão de séries temporais e Modelo Oculto de Markov. Novas técnicas de detecção de anomalia e predição de tráfego baseadas em aprendizado profundo são apresentadas, tais como redes neurais recorrentes, redes neurais com memória longa de curto prazo e redes neurais convolucionais. Por fim, são apresentados os desafios da próxima geração de rede móveis (6G), novos paradigmas e um estudo de caso com exercício prático de desenvolvimento de um exemplo de detecção de anomalia e predição de tráfego através de ferramentas livres de código aberto.

show abstract

Section: As Ferramentas Baseadas Em Modelos Estatísticosunclassified

Segurança em Redes 5G: Oportunidades e Desafios em Detecção de Anomalias e Predição de Tráfego Baseadas em Aprendizado de Máquina

Barbosa¹,

Bezerra²,

Medeiros³

et al. 2021

Minicursos Do XXI Simpósio Brasileiro De Segurança Da Informação E De Sistemas Computacionais

View full text Add to dashboard Cite

show abstract

“…Yet, several challenges still present for the VANET's management and it's deployment in the instance of secured DC [14]. To secure communications between connected vehicles, Artificial Intelligence (AI) techniques, like [15], [16], [17], [18], are needed. The AI continuously utilizes the experience that it got in the augmentation of its cognitive capability concerning environment and also making the instant good decisions [19,20].…”

Section: Introductionmentioning

confidence: 99%

Improving The Security Of Data Communication in Vanets Using ASCII-ECC Algorithm

Sajini

Anita

Janet³

2021

Preprint

View full text Add to dashboard Cite

Now-a-days, with the augmenting accident statistics, Vehicular Ad-hoc Networks (VANET) are turning out to be more popular, which in-turn eradicates accidents in addition to damage to the vehicles together with populace. In a VANET, message can well be transmitted within a pre-stated region to attain safety of a system and also its efficacy. Next, it is challenge to ensure authenticity of messages in such a dynamic environment. Though some researchers have already worked on this, security has not been much focussed. Thus, secured data communication with enhanced security on the VANET environment utilizing the American Standard Code for Information Interchange centred Elliptic Curve Cryptography (ASCII-ECC) is proposed. The proposed scheme allows all vehicles to register with the Trusted Authority (TA) and uses Median-centred K-Means (MKM) to find out the cluster head among all cluster and at the last phase verification is done by TA. The performance of the proposed algorithm ASCII-ECC is analysed and proved that the system renders better performance when it is weighed against the top-notch methods.

show abstract

“…Two important research directions are commonly reported in traffic prediction literature [ 17 , 18 , 19 ]: one based on parametric models (or statistic-based) and one based on nonparametric (nonconventional or AI-based) models [ 17 , 20 ]. Lately, due to the increased quantity of data and the computation capability [ 14 ], most contributions have been focusing on the AI-based approach, more specifically, the machine learning/deep learning (support vector regression (SVR), k-nearest neighbor, Bayesian Network, random forest, convolutional neural networks CNN, recurrent neural network RNN, graph CNN, long-short-term-memory LSTM), reinforcement learning and transfer-learning methods [ 11 , 14 , 17 , 21 , 22 , 23 , 24 ].…”

Section: Introductionmentioning

confidence: 99%

An Hybrid Approach for Urban Traffic Prediction and Control in Smart Cities

Culita

Caramihai

Dumitrache

et al. 2020

Sensors

View full text Add to dashboard Cite

Smart cities are complex, socio-technological systems built as a strongly connected System of Systems, whose functioning is driven by human–machine interactions and whose ultimate goals are the well-being of their inhabitants. Consequently, controlling a smart city is an objective that may be achieved by using a specific framework that integrates algorithmic control, intelligent control, cognitive control and especially human reasoning and communication. Among the many functions of a smart city, intelligent transportation is one of the most important, with specific restrictions and a high level of dynamics. This paper focuses on the application of a neuro-inspired control framework for urban traffic as a component of a complex system. It is a proof of concept for a systemic integrative approach to the global problem of smart city management and integrates a previously designed urban traffic control architecture (for the city of Bucharest) with the actual purpose of ensuring its proactivity by means of traffic flow prediction. Analyses of requirements and methods for prediction are performed in order to determine the best way for fulfilling the perception function of the architecture with respect to the traffic control problem definition. A parametric method and an AI-based method are discussed in order to predict the traffic flow, both in the short and long term, based on real data. A brief comparative analysis of the prediction performances is also presented.

show abstract

Artificial intelligence-based vehicular traffic flow prediction methods for supporting intelligent transportation systems

Cited by 120 publications

References 88 publications

Segurança em Redes 5G: Oportunidades e Desafios em Detecção de Anomalias e Predição de Tráfego Baseadas em Aprendizado de Máquina

Segurança em Redes 5G: Oportunidades e Desafios em Detecção de Anomalias e Predição de Tráfego Baseadas em Aprendizado de Máquina

Improving The Security Of Data Communication in Vanets Using ASCII-ECC Algorithm

An Hybrid Approach for Urban Traffic Prediction and Control in Smart Cities

Contact Info

Product

Resources

About