Multi-Lane Short-Term Traffic Forecasting With Convolutional LSTM Network

Ma, Yixuan; Zhang, Zhenji; Ihler, Alexander T.

doi:10.1109/access.2020.2974575

Cited by 53 publications

(44 citation statements)

References 48 publications

(55 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A convolutional LSTM based multilane prediction algorithm 50 in which CNN used to extract the neighboring lanes and up/downstream routing patterns. LSTM used to predict the dynamic behavior of temporal features in the lane 51 .…”

Section: Related Workmentioning

confidence: 99%

An attention‐based deep learning model for traffic flow prediction using spatiotemporal features towards sustainable smart city

Vijayalakshmi¹,

Ramar²,

Jhanjhi

et al. 2020

Int J Communication

View full text Add to dashboard Cite

Summary In the development of smart cities, the intelligent transportation system (ITS) plays a major role. The dynamic and chaotic nature of the traffic information makes the accurate forecasting of traffic flow as a challengeable one in ITS. The volume of traffic data increases dramatically. We enter the epoch of big data. Hence, a 1deep architecture is necessary to process, analyze, and inference such a large volume of data. To develop a better traffic flow forecasting model, we proposed an attention‐based convolution neural network long short‐term memory (CNN‐LSTM), a multistep prediction model. The proposed scheme uses the spatial and time‐based details of the traffic data, which are extracted using CNN and LSTM networks to improve the model accuracy. The attention‐based model helps to identify the near term traffic details such as speed that is very important for predicting the future value of flow. The results show that our attention‐based CNN‐LSTM prediction model provides better accuracy in terms of prediction during weekdays and weekend days in the case of peak and nonpeak hours also. We used data from the largest traffic data set the California Department of Transportation (Caltrans) for our prediction work.

show abstract

Section: Related Workmentioning

confidence: 99%

An attention‐based deep learning model for traffic flow prediction using spatiotemporal features towards sustainable smart city

Vijayalakshmi¹,

Ramar²,

Jhanjhi

et al. 2020

Int J Communication

View full text Add to dashboard Cite

show abstract

“…In fact, spatio-temporal users' dynamics captured by SDN controllers are collected by the NFVO, which can consequently perform mobility prediction. Specifically, the Convolutional LSTM (ConvLSTM) architecture, which has been initially introduced for precipitation nowcasting [24] and recently investigated also for traffic forecasting [25], is adopted for this purpose. The ConvLSTM is a neural network based on LSTM [26], with the convolution operator as input, forget, and output gates instead of the elementwise or Hadamard product [24].…”

Section: B Recognition Of User Mobility Patternsmentioning

confidence: 99%

Towards an Optimal Management of the 5G Cloud-RAN through a Spatio-Temporal Prediction of Users’ Dynamics

Rago

Ventrella

Piro

et al. 2020

2020 Mediterranean Communication and Computer Networking Conference (MedComNet)

View full text Add to dashboard Cite

In the emerging 5G architecture, the Cloud-Radio Access Network (Cloud-RAN) offers the possibility to dynamically configure virtual resources and network functionalities very close to end-users, while jointly considering bandwidth, computing, latency, and memory capabilities requested by heterogeneous applications, the channel quality experienced by end-users, mobility, and any kind of system constraints. By capitalizing on recent scientific results and standardization activities on 5G, this short paper presents a preliminary design of an ETSI-NFV compliant architecture willing to support the implementation of advanced protocols, algorithms, and methodologies for the optimal management of the 5G Cloud-RAN. Its components and functionalities have been sketched by harmoniously integrating Software-Defined Networking (SDN) facilities, Multi-access Edge Computing (MEC), and deep learning. Herein, spatio-temporal users' dynamics are collected by SDN controllers and predicted by a high-level orchestrator through a Convolutional Long Short-Term Memory scheme. Then, the outcomes of the prediction process are adopted to dynamically configure the Cloud-RAN (i.e., by using any kind of customizable algorithm). Some of the capabilities of the proposed approach are preliminarily evaluated by considering the autonomous driving use case and real mobility traces. Moreover, the paper concludes by reporting an overview of future directions of this research activity.

show abstract

“…However, they have been unable to capture long-term dependencies. In recent years, recurrent neural network (RNN) and its derivative long short-term memory network (LSTM) [22], [23] and gated recurrent neural network [24] show great advantages in terms of sequence prediction tasks. Among them, the network of long short-term memory has achieved good results in petrochemical field [25], transportation field [26], [27] and medical field [28], [29].…”

Section: B the Establishment Of Prediction Modelsmentioning

confidence: 99%

Smart Online Fuel Sulfur Prediction in Diesel Hydrodesulfurization Process

Huang

et al. 2020

IEEE Access

View full text Add to dashboard Cite

In the process industry, online operators usually adjust the entire production process in time according to changes in certain key indicators. However, it is difficult to directly obtain the value of key indicators due to the complex structure and the numerous processes of the process industry. This paper proposes a method to achieve high precision on-line prediction of key indicators in the process industry. The method first performs data preprocessing of multi-source heterogeneous time series data involved in industrial processes based on the professional knowledge, which not only keeps the prediction error within 10%, but also reduces the prediction time. Then a model framework is constructed based on LSTM neural network and the error correction algorithm is proposed to improve prediction accuracy based on real-time error, which directly causes the error to drop by 3%-5%. At the end, a set of multi-mode online training strategies and related trigger conditions are designed to perform predicting online. Diesel hydrodesulfurization is a typical case in the process industry. The effectiveness of the proposed method is empirically studied by applying its actual data sets. Through the comparison with other traditional wellknown forecasting models, and models optimized by adjusting parameters, the experimental results demonstrate that the method can achieve the great prediction performance in terms of both accuracy and stability.

show abstract

Multi-Lane Short-Term Traffic Forecasting With Convolutional LSTM Network

Cited by 53 publications

References 48 publications

An attention‐based deep learning model for traffic flow prediction using spatiotemporal features towards sustainable smart city

An attention‐based deep learning model for traffic flow prediction using spatiotemporal features towards sustainable smart city

Towards an Optimal Management of the 5G Cloud-RAN through a Spatio-Temporal Prediction of Users’ Dynamics

Smart Online Fuel Sulfur Prediction in Diesel Hydrodesulfurization Process

Contact Info

Product

Resources

About