MA trip generation

Overview

This project aims to develop Massachusetts-specific trip generation models for land use projects. The primary focus is on utilizing location-based service (LBS) data to create accurate and efficient trip generation models, particularly for urban sites that benefit from proximity to public transportation. (Funding agency: MassDOT)

Outputs

• Final Report from mass.gov
• TRB RIP Link
• Research project statement

Sample output showing vehicle count in specified direction.

Project Duration

March 2021 - September 2023

Video Data Collection and Processing

• Objective: Develop Massachusetts-specific trip generation models using LBS data.
• Methodology: Utilize YOLOv8 Medium for object detection and tracking in RGB and IR traffic videos.
• Significance: Overcome limitations of traditional data collection methods, particularly for urban sites.
• Procedure developed for video data collection using trail cameras and automatic processing using AI algorithms.
• RGB and IR videos tested for vehicle detection and tracking.
• YOLOv8-Medium adopted for vehicle detection, with Deep-SORT algorithm for tracking.
• Results showed the feasibility and potential of using LBS data, with lower absolute percentage error compared to ITE estimates.

Videos were collected from over 25 locations. Several thousand video clips were collected and processed. Training and inference was done using A100 and RTX 4090 GPUs.

Comparision of RGB and IR camera image quality.

Comparision of detections in RGB and IR camera videos at night.

---

Deep Learning model

• The YOLOv8 model was chosen for its balance between speed and accuracy in the context of the project.
• YOLOv8-Medium was used in the experiments, providing optimal results in terms of inference speed and model accuracy.
• Larger YOLOv8 models did not significantly outperform YOLOv8-Medium.

Link

Bureau of Transportation Statistics- https://rosap.ntl.bts.gov/view/dot/72776

Title : Developing Massachusetts-Specific Trip Generation Models for Land Use Projects
Creator(s) : Xie, Yuanchang;Loesch, Brandon;Bhuyan, Zubin;Logozzo, Francesco;Chen, Danjue;Liu, Benyuan (Ben);
Corporate Creator(s) : University of Massachusetts at Lowell
Corporate Contributor(s) : Massachusetts. Dept. of Transportation. Office of Transportation Planning;United States. Department of Transportation. Federal Highway Administration;
Contributor(s) : Lucien, Lionel
Published Date : 2023-09-01
Report Number : 23-046
URL : https://rosap.ntl.bts.gov/view/dot/72776