Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Accounting for Work Zone Disruptions in Traffic Flow Forecasting

Yuanjie Lu, Amarda Shehu, David Lattanzi

TL;DR

A novel ``Graph Convolutional Network for Roadway Work Zones" model that includes a novel data fusion mechanism and a new heterogeneous graph aggregation methodology to accommodate work zone information in spatio-temporal dependencies among traffic states is presented.

Abstract

Traffic speed forecasting is an important task in intelligent transportation system management. The objective of much of the current computational research is to minimize the difference between predicted and actual speeds, but information modalities other than speed priors are largely not taken into account. In particular, though state of the art performance is achieved on speed forecasting with graph neural network methods, these methods do not incorporate information on roadway maintenance work zones and their impacts on predicted traffic flows; yet, the impacts of construction work zones are of significant interest to roadway management agencies, because they translate to impacts on the local economy and public well-being. In this paper, we build over the convolutional graph neural network architecture and present a novel ``Graph Convolutional Network for Roadway Work Zones" model that includes a novel data fusion mechanism and a new heterogeneous graph aggregation methodology to accommodate work zone information in spatio-temporal dependencies among traffic states. The model is evaluated on two data sets that capture traffic flows in the presence of work zones in the Commonwealth of Virginia. Extensive comparative evaluation and ablation studies show that the proposed model can capture complex and nonlinear spatio-temporal relationships across a transportation corridor, outperforming baseline models, particularly when predicting traffic flow during a workzone event.

Accounting for Work Zone Disruptions in Traffic Flow Forecasting

TL;DR

A novel ``Graph Convolutional Network for Roadway Work Zones" model that includes a novel data fusion mechanism and a new heterogeneous graph aggregation methodology to accommodate work zone information in spatio-temporal dependencies among traffic states is presented.

Abstract

Traffic speed forecasting is an important task in intelligent transportation system management. The objective of much of the current computational research is to minimize the difference between predicted and actual speeds, but information modalities other than speed priors are largely not taken into account. In particular, though state of the art performance is achieved on speed forecasting with graph neural network methods, these methods do not incorporate information on roadway maintenance work zones and their impacts on predicted traffic flows; yet, the impacts of construction work zones are of significant interest to roadway management agencies, because they translate to impacts on the local economy and public well-being. In this paper, we build over the convolutional graph neural network architecture and present a novel ``Graph Convolutional Network for Roadway Work Zones" model that includes a novel data fusion mechanism and a new heterogeneous graph aggregation methodology to accommodate work zone information in spatio-temporal dependencies among traffic states. The model is evaluated on two data sets that capture traffic flows in the presence of work zones in the Commonwealth of Virginia. Extensive comparative evaluation and ablation studies show that the proposed model can capture complex and nonlinear spatio-temporal relationships across a transportation corridor, outperforming baseline models, particularly when predicting traffic flow during a workzone event.
Paper Structure (25 sections, 6 equations, 5 figures, 8 tables, 1 algorithm)

This paper contains 25 sections, 6 equations, 5 figures, 8 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. GCN-RWZ: Architecture and Methodology
  4. Road Network as Graph
  5. Heterogeneous Graph Aggregation
  6. Speed feature map $X^{S}$
  7. Average history speed feature map $X^{AS}$
  8. Diff feature Map $X^{D}$
  9. Time matrix
  10. Construction feature map $X^{C}$
  11. Feature aggregation
  12. Hypergraph convolution network
  13. Multi-head Spatial-Temporal Attention Mechanism
  14. Spatial hypergraph convolution network
  15. Temporal convolutional
  16. ...and 10 more sections

Figures (5)

  • Figure 1: Overall framework of the developed speed prediction methodology and of the proposed GCN-RWZ model.
  • Figure 2: Left: The Richmond data set is derived from an in-roadway sensor network in Richmond, Virginia. Right: The Tyson's data set is derived from RITIS data in the Tyson's Corner region of Northern Virginia.
  • Figure 3: Predictive accuracy during workzone disruptions where the ground truth change in speed relative to historical average is greater than $\pm 5$ MPH. Forecast shown for 3, 6, and 12 time-step predictions. Left: Richmond data set (15-minute sample interval). Right: Tyson's data set (5-minute sample interval)
  • Figure 4: RMSE heatmap representation of GCN-RWZ (left) and DGCRN (right) at forecast intervals (x-axis) over 8 random road segments (y-axis), Richmond data set under work zone disruption.
  • Figure 5: Relative speed forecasting accuracy with 90 mins of forecast length on three road segments under the impact of the construction work from 04/03/2019 to 04/05/2019: (a) Road segment = "40488"; (b) Road segment = "140010" and from 02/13/2019 to 02/15/2019: (c) Road segment = "140052";(d) Road segment = "140315".