Layout Option

Master of Science in Engineering

Video and Acoustic Analysis of Vehicle-Pedestrian Impacts

Date of Award

2024

Document Type

Thesis

Degree Name

Engineering

Department

Engineering

First Advisor

Dr. Theresa Atkinson

Second Advisor

Dr. Massoud Tavakoli

Third Advisor

Dr. Daniel Ludwigsen

Abstract

Physical evidence deposited by vehicles as a result of an accident may not always be fully documented by the investigating law enforcement officers by means of measurements. Many times, the investigating officers are able to document the location of the vehicles at rest, but can only measure and locate some of the related tire marks and/or gouges that may be present on the roadway as a result of the collision. The officers may further supplement their investigation by photographing the scene of the accident. The photographs often depict the position and geometry of other physical evidence relative to the roadway or other geographic features. However, the forensic engineer may require more detailed information regarding the physical evidence than is shown in the photographs or measured by the officers, and must therefore take further steps to extract it. Depending on the evaluation required in the case, the forensic engineer would need to further analyze the accident by an inspection of the accident site and/or vehicles. After collecting the available information from the measurements made by the officers, the photographs taken at the scene, and the inspections by the forensic engineer, the data must then be compiled in a comprehensive form to further evaluate the dynamics of the accident.

Modern vehicles record dynamic data from several on-board sensors for events that could precede a crash. These data can be used to reconstruct the behavior of a vehicle, although the accuracy of these reconstructions has not yet been quantified. Here, we evaluated various methods of reconstructing the vehicle kinematics in vehicle-pedestrian impacts for three different vehicles – 2003 Pontiac Vibe, 2005 Mercury Montego and 2009 Jeep Patriot based on video analysis. The test dummy a.k.a. Anthropomorphic Test Device (ATD) was stationary and the vehicles were made to impact them at speeds varying between 25-40 mph (40-64 Kmph) in nine different impacts tests- three tests per vehicle mentioned above.

Audio (MP3) files were recorded for these vehicle-pedestrian impacts using multiple GoPro Hero8 cameras and were later analyzed to find corresponding data correlations between the audio and video files based on level vs. time analysis. A comprehensive analysis and comparison of the audio recording capabilities between GoPro Hero8, a free field microphone and a binaural head have also been provided.

Link to Full Text

Share

COinS