Data mining for modeling drivers’ responses to cutting-in and out of traffic – UROP Spring Symposium 2021

Data mining for modeling drivers’ responses to cutting-in and out of traffic

Jim Vega

Jim Vega

Pronouns: He/Him/His

Research Mentor(s): Brian Lin, Assistant Research Scientist
Research Mentor School/College/Department: University of Michigan Transportation Research Institute, College of Engineering
Presentation Date: Thursday, April 22, 2021
Session: Session 2 (11am – 11:50am)
Breakout Room: Room 18
Presenter: 3

Event Link

Abstract

Understanding drivers’ maneuvers can help improve the technology that assists drivers, from the perspectives of user-friendliness, acceptability, and safety. In the past few years, UMTRI has applied driving data sets to conduct studies for the investigation of driver’s behavior and vehicle control characteristics: use of turn signal at intersections (Sullivan et al., 2015) and on the highway (Lin & Bao, 2019), distraction (Li et al., 2018; Wang et al., 2017), aggressive driving (Feng et al., 2017), and interaction with bicyclists (Feng et al., 2018), etc. Mining these naturalistic data sets provides great objective evidence to infer drivers’ behaviors. This project aims to investigate US drivers’ maneuver profiles in the scenario of having a leading vehicle cutting in and out of the lane on highways, by analyzing information extracted from UMTRI’s SPMD (Safety Pilot Model Deployment) database. To tackle the research questions, we collaboratively focused on extracting and validating the desired samples from the database using SQL. Finalizing our querying algorithm, we found an approximate 80% accuracy rate on the validations we made through the video viewer. Current steps in our approach involves analysis and categorization of the information we have in order to formulate a solid conclusion of our findings.

Authors: Jim Vega, Jason Hu, Brian Lin
Research Method: Computer Programming

lsa logoum logo