skip to main content

<< Return to Webinar Files

Webinar Presentation

HTML version of the presentation
Image descriptions are contained in brackets. [ ]

A Digital Twin for Traffic Monitoring & Proactive Incident Management
(May 11, 2021)

Estimating Impacts of Traffic Crashes
Presenter: Qifan Nie, PhD
Presenter’s Org: College of Engineering, University of Alabama

T3 and T3e webinars are brought to you by the Intelligent Transportation Systems (ITS) Professional Capacity Building (PCB) Program of the U.S. Department of Transportation (USDOT)’s ITS Joint Program Office (JPO). References in this webinar to any specific commercial products, processes, or services, or the use of any trade, firm, or corporation name is for the information and convenience of the public, and does not constitute endorsement, recommendation, or favoring by the USDOT.


[The slides in this presentation contain the University of Alabama’s logo.]

Slide 1: Student Presentation 4

Title: Estimating Impacts of Traffic Crashes

Students: Qifan Nie, Postdoctoral Researcher; Zihe Zhang, PhD Candidate;

Advisors: Jun Liu, PhD, Assistant Professor; Alex Hainen, PhD, Associate Professor

  • The objectives of the research are to:
    • Identify the correlates of spatiotemporal impacts of traffic crashes
    • Develop models to estimate the spatiotemporal impacts of traffic crashes on traffic flow

Key Work:

  • Identification of the spatiotemporal extent of traffic congestion/queue
  • Measure of impacts
  • Machine learning to identify correlates of impacts

[This slide contains two images: (1) a photo of two lanes of cars in traffic and (2) a matrix.]

Slide 2: The Same Data…

HERE Traffic Data

  • Speed data

ALGO Incident Data

  • Incident information

CARE Crash Data

  • Tabulated crash report data

Highway Performance Monitoring System (HPMS) Data

  • Freeway segments and ramps characteristics

[This slide contains four logos: (1) the logo of the Alabama Department of Transportation, (2) the logo of HERE, (3) the logo of ALGO, and (4) the logo of CARE.]

Slide 3: Measuring Spatiotemporal Impacts

  • Maximum Queue Length
  • Time at Maximum Queue Length

[This slide contains a graph of the time and distance related to major crashes at mile marker 256.]

Slide 4: Measuring Spatiotemporal Impacts

  • Volume ‐ Spatiotemporal Extent of a Queue
    • Speed reduction × Segment length × Time window

[This slide contains a three dimensional graph of the spatiotemporal extent of a queue. The y‐axis is labeled “Segment,” the x‐axis is labeled “Time Window,” and the z‐axis is labeled “Speed Reduction.”]

Slide 5: Crash Impact Modeling

  • Maximum Queue Length
    • To estimate the maximum queue length given a set of variables (traffic, roadways, crash characteristics)
  • Time at Maximum Queue Length
    • To estimate the time at maximum queue length after a crash given a set of variables (traffic, roadways, crash characteristics)
  • Volume (spatiotemporal extent of queue)
    • To estimate the spatiotemporal impacts of a crash given a set of variables (traffic, roadways, crash characteristics)

[This slide contains a background of an aerial image showing several levels of roads with cars on them.]

Slide 6: Variables

HPMS and Incident‐related variables

Variable Name Description
Volume Mean of total speed drops Dependent variables
Longest queue Longest queue length within 60 mins post‐crash Dependent variables
Time longest queue Timestamp of the longest queue length Dependent variables
Roadway clearance time Time to clear the roadway Independent variables
Incident clearance time Time to clear the incident/crash Independent variables
Initial Severity Incident/crash initial severity Independent variables
Weekday Weekday, weekend Independent variables
Crash type Crash type Independent variables
RampDis Distance to the nearest ramp Independent variables
AADT AADT for location (Ramp AADT) Independent variables
Nlanes Number of lanes Independent variables
Ramp Ramp or not Independent variables
Urban Rural/Urban Independent variables
AccessCon Access control Independent variables
Time of day a.m. peak (06:00–10:00), daytime (10:00–16:00), p.m. peak (16:00–20:00), night (20:00–06:00) Independent variables

Slide 7: Maximum Queue Length

[This slide contains two graphs: (1) a bar graph titled “Distribution of the longest queue length” with “Count” on the y‐axis and “Queue (Unit: mile)” on the x‐axis. (2) a pie chart titled “Queue Length” with the sections of the chart being “Shorter than 1 mile” at 8.11%, “1–2 miles” at 8.35%, “2–3 miles” at 17.55%, “3–4 miles” at 30.02%, and “Longer than 4 miles” at 35.96%.]

Slide 8: Time at Maximum Queue Length

[This slide contains two graphs: (1) a bar graph titled “Distribution of Time at Maximum Queue Length” with “Count” on the y‐axis and “Time at Maximum queue length (Unit: minute)” on the x‐axis. (2) a pie chart titled “Maximum queue length” with the sections of the chart being “45–60 minutes” at 18.20%, “30–45 minutes” at 20.09%, “15–30 minutes” at 21.40%, and “less than 15 minutes” at 40.32%.]

Slide 9: Volume (Speed Reduction × Segment length × Time window)

[This slide contains two graphs: (1) a bar graph titled “Distribution of the volume” with “Count” on the y‐axis and “Volume” on the x‐axis. (2) a pie chart titled “Volume” with the sections of the chart being “15,000–20,000” at 4.94%, “10,000–15,000” at 22.47%, “5,000–10,000” at 32.59%, and “less than 5,000” at 40.00%.]

Slide 10: Independent Variables

  • Roadway clearance time
  • Incident clearance time

[This slide contains two charts: (1) a pie chart titled “Roadway clearance time” with the sections of the chart being “0–15 minutes” at 14.65%, “15–30 minutes” at 17.19%, “30–45 minutes” at 14.16%, and “longer than 45 minutes” at 54.00%. (2) a pie chart titled “Incident clearance time” with the sections of the chart being “0–15 minutes” at 9.56%, “15–30 minutes” at 15.86%, “30–45 minutes” at 15.25%, and “longer than 45 minutes” at 59.32%.]

Slide 11: Independent Variables

  • Initial Severity
  • Crash Type

[This slide contains two charts: (1) a pie chart titled “Initial Severity” with the sections of the chart being “Extremely High” at 2.54%, “High” at 5.57%, “Medium” at 24.46%, and “Low” at 67.43%. (2) a pie chart titled “Crash Type” with the sections of the chart being “Overturned vehicle” at 3.27%, “Major crash” at 12.83%, “Moderate crash” at 26.76%, and “Minor crash” at 57.14%.]

Slide 12: Modeling Methods

  • Logistic Regression
  • Random Forest

[This slide contains two charts: (1) a line graph of logistic regression and (2) a flow chart of test sample input.]

Slide 13: Modeling Results ‐ Maximum Queue Length

[This slide contains two matrices: (1) a confusion matrix titled “Logistic Regression” and (2) a confusion matrix titled “Random Forest.”]

Slide 14: Modeling Results ‐ Maximum Queue Length

Feature importance (permutation importance)

  • Describes which features are relevant
  • Provides a highly compressed, global insight into the model’s behavior
Variable Name Description
Near Dist FT distance to the nearest ramp
AADT AADT for location (Ramp AADT)
ThroughLa number of lanes
Time of day a.m. peak (06:00–10:00),
daytime (10:00–16:00),
p.m. peak (16:00–20:00),
night (20:00–06:00)
Weekday weekdays, weekend
Crash type Crash type

[This slide contains a bar graph titled “Visualizing Import Features” with the y‐axis labeled “Features” and the x‐axis labeled “Feature Importance Score.” The six features in the table are highlighted in the bar graph and have the highest feature importance score.]

Slide 15: Modeling Results ‐ Time at Maximum Queue Length

[This slide contains two matrices: (1) a confusion matrix titled “Logistic Regression” and (2) a confusion matrix titled “Random Forest.”]

Slide 16: Modeling Results ‐ Time at Maximum Queue Length

Feature importance (permutation importance)

Variable Name Description
Near Dist FT distance to the nearest ramp
AADT AADT for location (Ramp AADT)
ThroughLa number of lanes
Time of day a.m. peak (06:00–10:00),
daytime (10:00–16:00),
p.m. peak (16:00–20:00),
night (20:00–06:00)
Weekday weekdays, weekend
Crash type Crash type

[This slide contains a bar graph titled “Visualizing Import Features” with the y‐axis labeled “Features” and the x‐axis labeled “Feature Importance Score.” The six features in the table are highlighted in the bar graph and have the highest feature importance score.]

Slide 17: Modeling Results ‐ Volume (Speed drop × Segment length × Time window)

[This slide contains two matrices: (1) a confusion matrix titled “Logistic Regression” and (2) a confusion matrix titled “Random Forest.”]

Slide 18: Modeling Results ‐ Volume

Feature importance (permutation importance)

Variable Name Description
Near Dist FT distance to the nearest ramp
AADT AADT for location (Ramp AADT)
ThroughLa number of lanes
Time of day a.m. peak (06:00–10:00),
daytime (10:00–16:00),
p.m. peak (16:00–20:00),
night (20:00–06:00)
Weekday weekdays, weekend
Urban Rural/Urban

[This slide contains a bar graph titled “Visualizing Import Features” with the y‐axis labeled “Features” and the x‐axis labeled “Feature Importance Score.” The six features in the table are highlighted in the bar graph and have the highest feature importance score.]

↑ Return to top

SUPPORT

Technical Assistance is available to Federal, State and local transportation agencies through:

ITS Peer Program - The ITS Peer-to-Peer Program puts you in touch with technical experts or experienced peers.

ITS Help Line - The ITS Help Line provides technical support by email or telephone at 866-367-7487.

STAY CONNECTED

go to twitter    go to Facebook    go to Instagram    go to Linkedin    go to email