Spatio-Temporal AutoEncoder for Traffic Flow Prediction

Article Engineering, Civil

Traffic Demand Prediction Based on Dynamic Transition Convolutional Neural Network

Bowen Du et al.

Summary: A Dynamic Transition Convolutional Neural Network (DTCNN) is proposed for precise traffic demand prediction, which incorporates spatial distributions, dynamics of demand, and environmental factors. By constructing a transition network and designing dynamic transition convolution units, the method effectively handles the challenges in traffic demand prediction, as validated by experiments on NYC taxi and bike-sharing data.

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS (2021)

添加到收藏夹

Review Multidisciplinary Sciences

Time-series forecasting with deep learning: a survey

Bryan Lim et al.

Summary: This article explores the diversity of time-series datasets across different domains, discussing common encoder and decoder designs, and how deep learning models incorporate temporal information into predictions. Additionally, recent developments in hybrid deep learning models are highlighted, along with ways in which deep learning can facilitate decision support.

PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES (2021)

添加到收藏夹

Article Environmental Studies

Optimizing electric bus charging infrastructure considering power matching and seasonality

Xiaohan Liu et al.

Summary: This research introduces a novel optimization model for electric bus charging station location, charger configuration, charging time, and vehicle flow considering power matching and seasonality. A surrogate-based optimization approach is used to efficiently solve the mixed integer nonlinear program, revealing significant performance differences in vehicle scheduling and charging among different bus fleets in a Beijing-based transit system. Interesting findings on the distribution of vehicle flows for charging provide strong evidence to consider power matching in the bus charging infrastructure layout.

TRANSPORTATION RESEARCH PART D-TRANSPORT AND ENVIRONMENT (2021)

添加到收藏夹

Article Economics

Improving flex-route transit services with modular autonomous vehicles

Xiaohan Liu et al.

Summary: This paper introduces a novel operational design for flexible route transit services with MAVs to reduce operation costs and enhance customer service quality. A two-stage solution framework utilizing dynamic programming and heuristic methods is developed to address bus scheduling and ride-sharing problems efficiently. Numerical examples and case studies demonstrate the effectiveness of the proposed operational design.

TRANSPORTATION RESEARCH PART E-LOGISTICS AND TRANSPORTATION REVIEW (2021)

添加到收藏夹

Article Engineering, Civil

Forecasting Transportation Network Speed Using Deep Capsule Networks With Nested LSTM Models

Xiaolei Ma et al.

Summary: The study proposes a new framework for network-level traffic forecasting using CapsNet and NLSTM, demonstrating their superiority in capturing complex spatiotemporal traffic patterns through visualizing and quantitatively evaluating experimental results.

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

Dynamic and Multi-faceted Spatio-temporal Deep Learning for Traffic Speed Forecasting

Liangzhe Han et al.

Summary: This paper explores the use of Dynamic Graph Neural Networks for traffic speed forecasting, presenting a dynamic graph construction method, dynamic graph convolution module, and multi-faceted fusion module to better capture the spatio-temporal characteristics of road segments and achieve state-of-the-art prediction performances by incorporating information from traffic volumes and speeds.

KDD '21: PROCEEDINGS OF THE 27TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING (2021)

添加到收藏夹

Article Engineering, Civil