Heavy Truck Count Study

Analyzing Heavy Truck Traffic on West Saint Catherine Street with the help of AI Object detection

Introduction

The great weight of heavy trucks accelerates wear and damage to neighborhood street surfaces. This is especially true in Old Louisville and Limerick where many of the streets and structures were constructed in the late 19th century. Heavy trucks also emit high levels of particulate and noise pollution, and are dangerous to pedestrians and cyclists due to their poor visibility and long stopping distances.

An eighteen-wheeler traffic jam
Trucks fill the street while the person using a bike for transport is relegated to the sidewalk

Truck Types

Commercial trucks in the US are broadly classified as either Heavy, Medium, or Light duty based on the Gross Vehicle Weight Rating (GVWR). We can infer which category a truck belongs to based on its axle configuration.

Heavy Duty Trucks

Medium and Light Duty Trucks

Methods

For the week of Monday, November 4th – Friday, November 8th, 2024, we monitored heavy truck traffic on the 500 block of West Saint Catherine St. AI object detection was applied to the video feed of a camera with a view of the street using Frigate NVR. Frigate was configured to save any video recordings where trucks were detected for review.

For this initial investigation, we used a general AI object classification model which does not include any ability to distinguish different types of truck. The video footage containing trucks was manually reviewed and the truck type was recorded. The business and truck ID number were also recorded if visible.

Another limitation was that the object detection did not perform well outside of daylight hours due to lower-quality images.

If future data collection is needed, we will work on building a more specific model which can do these tasks automatically.

Results

Results showed significant amounts of truck traffic, an average of 142.8 trucks per day during the study period. Heavy Duty trucks accounted for 52.4, of which 17.2 were semi tractor-trailers. Below we have a summary table and some interactive graphs of the complete data set.

Average Weekday Traffic

Daylight hours only (approx. 6:45 AM to 6:00 PM)

Truck Class Trucks per day
Semi Tractor-Trailer 17.2
Semi Tractor Only 1.0
Dual Axle 32.8
Tri Axle 1.4
Heavy Duty Total 52.4
Single Axle, Dual Rear Wheel 67.2
Single Axle, Single Rear Wheel 23.2
Light and Medium Duty Total 90.4
Grand Total 142.8

Hourly Trends

Peak truck traffic occurred from 8:00 AM to 1:00 PM, with over 16 trucks per hour and an average of 2.5 semi tractor-trailers per hour.

Truck Operators

Where possible, the company or organization listed on the truck was recorded. Below are all operators with 6 or more trips during the week study period.

Three of the top four, and eight of the total are waste disposal companies, predominately operating heavy duty dual axle trucks. Most likely, these operators are using the neighborhood as a cut-through between I-65 to the east and the Rumpke transfer station to the west. Simpson Agri, the most frequent Semi Truck operator, is probably also using the neighborhood as a cut-through from I-65 to deliver grain to the Brown-Forman distillery on Dixie Highway

Truck ID numbers were also recorded where possible and Any Time Waste Systems truck No. 31 made the most trips of any truck with 14 for the week!

What’s Next?

Leverage this data

  • Support heavy truck ordinance -> remove neighborhood streets from Trucking GPS Routes
  • Work with businesses, public partners to find alternate routes and reduce truck trips

Further Investigation

  • Similarly asses truck traffic in other streets/neighborhoods
  • Fully automate truck ID and classification
  • Estimate vehicle speeds
  • Collect noise and pollution data in sync with truck count

Acknowledgments

The idea for this project was largely inspired by a very cool truck-counting project by the Center for Neighborhood Technology in Chicago. Check it out at: Chicago Truck Data Portal

Free and Open Source Software was instrumental in being able to collect and analyze the data. Especially:

And lastly a big thank you to all the Old Louisville and Limerick neighbors who shared their encouragement, enthusiasm, and feedback!

News Coverage