Integrated Modeling for Road Condition Prediction Phase 3 Project Report

Printable version [PDF 2.0 MB]
You may need the Adobe® Reader® to view the PDFs on this page.

U.S. Department of Transportation
Federal Highway Administration
Office of Operations
1200 New Jersey Avenue, SE
Washington, DC 20590
ops.fhwa.dot.gov

October 2020
FHWA-HOP-20-061

---

Table of Contents

[ Notice and Quality Assurance Statement ] [ Technical Report Documentation Page ] [ SI Modern Metric Conversion Factors ] [ List of Acronyms ]

Executive Summary
Chapter 1. Introduction
	Background
	Purpose
	Scope
	Document Overview
Chapter 2. Project Description
	Methodology and Approach
	Project Tasks and Deliverables
	Stakeholder Engagement
Chapter 3. Implementation and Deployment
	User Needs
	Application Scenarios
		Variable Speed Limits
		Enhanced Traveler Information
		Enhanced Intelligent Signal Controls
		Maintenance
		Freight
		Work Zones
		Travelers
		Emergency Response
	System Description
		Data Collection
		Forecast Model Components
		Data Store
		User Interface
	Study Area Description and Modeling
	Traffic Estimation and Prediction System Traffic Model
		Traffic Flow Model
		Weather Adjustment Factors
		Online Traffic Flow Model Update
		Time-Dependent Origin-Destination Matrix
		Offline Calibration of the Origin-Destination Matrix
		Online Calibration of the Origin-Destination Matrix
	Machine Learning-Based Traffic Prediction
		Model
		Data
		Model Calibration
		Results
Chapter 4. Evaluation
	Introduction
	Summary of Findings
		Did Integrated Modeling for Road Condition Prediction Have an Operational Impact?
		Did Users Consider Integrated Modeling for Road Condition Prediction Information as Useful?
Chapter 5. Analysis, Conclusions, and Recommendations for Further Study
	Lessons Learned
	Deployment Considerations
	Potential Applications
		Pavement State Prediction
		Travel Time Predictions
		Flooding Events
	Conclusions
	Recommendations for Further Study, Development, and Application
Bibliography

List of Figures

Figure 1. Diagram. Components of the Integrated Modeling for Road Condition Prediction.

Figure 2. Screenshot. User interface of an example map in the Integrated Modeling for Road Condition Prediction system.

Figure 3. Map. Kansas City metropolitan study area of the Integrated Modeling for Road Condition Prediction.

Figure 4. Map. Selected detectors for traffic flow model calibration by identifier code.

Figure 5. Graph. Sensitivity analysis of different weights.

Figure 6. Diagram. Framework of online traffic demand calibration in Traffic Estimation and Prediction System.

Figure 7. Diagram. Proposed machine learning-based prediction model algorithm.

Figure 8. Diagram. Representation of the Integrated Modeling for Road Condition Prediction machine learning-based prediction (MLP) process.

Figure 9. Graph. Prediction of normal case with daily traffic patterns.

Figure 10. Graphs. Predictions of speeds with incident on link.

Figure 11. Graphs. Predictions of speeds with rain and snow weather condition.

Figure 12. Graph. Examples of predictions for special events.

Figure 13. Graphs. Examples of predictions for links without detectors.

Figure 14. Illustration. Integrated Modeling for Road Condition Prediction relationship to other weather-responsive management strategy tools.

Figure 15. Screenshot. Pavement state example — 4:30 a.m., January 22, 2020.

Figure 16. Screenshot. Pavement state prediction example — 7 a.m., January 22, 2020.

Figure 17. Screenshot. Route travel time early in a winter storm.

Figure 18. Screenshot. Route travel time prediction for a winter storm.

Figure 19. Screenshot. Hydrological event alert for July 27, 2017.

Figure 20. Screenshot. Local hydrological event map for July 27, 2017.

List of Tables

Table 1. Characteristics of traffic count data.