Wide-Attention and Deep-Composite Model for Traffic Flow Prediction in Transportation Cyber–Physical Systems

Abstract

Recently, traffic flow prediction has drawn significant attention because it is a prerequisite in intelligent transportation management in urban informatics. The massively-available traffic data collected from various sensors in Transportation Cyber-Physical Systems brings the opportunities in accurately forecasting traffic trend. Recent advances in deep learning shows the effectiveness on traffic flow prediction though most of them only demonstrate the superior performance on traffic data from a single type of vehicular carriers (e.g., cars) and does not perform well in other types of vehicles. To fill this gap, we propose a wide-attention and deep-composite (WADC) model consisting of a wide-attention module and a deep-composite module in this paper. In particular, the wide-attention module can extract global key features from traffic flows via a linear model with selfattention mechanism. The deep-composite module can generalize local key features via Convolutional Neural Network component and Long Short-Term Memory Network component. We also perform extensive experiments on different types of traffic flow datasets to investigate the performance of WADC model. Our experimental results exhibit that WADC model outperforms other existing approaches.

Dual Memory LSTM with Dual Attention Neural Network for Spatiotemporal Prediction

“…This model analyzes the correlations between variables and the relationships between historical samples and present samples. A wide-attention module and the deep-composite module are utilized in [ 22 ] to extract global key features and local key features. However, these methods [ 19 , 20 , 21 , 22 ] depend on local representations to some extent, which cannot get excellent performance on prediction task.…”

“…A wide-attention module and the deep-composite module are utilized in [ 22 ] to extract global key features and local key features. However, these methods [ 19 , 20 , 21 , 22 ] depend on local representations to some extent, which cannot get excellent performance on prediction task. An artificial neural network [ 23 ] has been proposed to model the unique properties of spatiotemporal data and derives a more powerful modeling capability to spatiotemporal data.…”

VF‐EFENet: A novel method for environmental sound filtering and feature extraction

“…Multidimensional features include basic customer information, behavior information, social information, and environmental information. The features of customers involved in gang fraud may be highly aggregated, so customers involved in gang fraud can be identified through group association 7–9 . However, Gang fraud is highly organized and concealed.…”

“…The features of customers involved in gang fraud may be highly aggregated, so customers involved in gang fraud can be identified through group association. [7][8][9] However, Gang fraud is highly organized and concealed. Financial companies need more data to detect gang fraud.…”

Stealthy switching attacks on sensors against state estimation in cyber‐physical systems

“…Cyber‐physical systems (CPSs) have attracted extensive attention since they play key roles in smart grid, 1 intelligent transportation, 2 healthcare systems 3 and so forth. The information and data in CPSs are transmitted through networks, which may lead to security risks.…”