Back to search

Georgia Institute of Technology

Impact of Street Lighting on Pedestrian Path Routing - Impedance Factor Implementation

Abstract

dc:description.abstract

Street lighting plays a crucial role in pedestrian safety and security during nighttime, when a significant portion of pedestrian fatalities occur despite low-traffic conditions. Studies show that improved street lighting increases pedestrian and obstacle visibility, reduces pedestrian-vehicle crashes, and significantly decreases crime rates. Lighting enhancements not only improve real safety, but also perceived safety, resulting in increased usage of pedestrian facilities. Studies on pedestrian route choice highlight the importance of street lighting, as evidenced by a 15 % decrease in pedestrian traffic when transitioning from bright white to dimmer yellow light along a particular route (Bystryantseva et al., 2020). Current pedestrian navigation systems utilize the shortest network travel time as the primary factor in providing shortest-path route recommendations. However, it is essential to incorporate other infrastructure factors that affect route choice, such as path grade or the presence/absence of ADA ramps, into the pedestrian network analysis to provide the best routes for specific users. The ST-CTN project under the ITS4US deployment program (FHWA, 2024) is incorporating pedestrian network impedance into the navigational system G-MAP, which would provide route recommendations that more realistically represent the factors affecting route choice. In this study, the impact of street lighting on pedestrian impedance has been assessed and is now recommended for use in nighttime route recommendations for pedestrian navigation. The initial step in implementing streetlight impedance values involved the development of a streetlight location dataset. This was accomplished using object detection techniques and applying transfer learning strategies using the YOLO v8 machine learning model. The wheelchair video dataset collected as a part of the ITS4US project has been annotated to train the model to a custom dataset of streetlights. A total of 3520 images were employed. The image dataset was split into training (60 %), validation (20 %), and testing (20 %) datasets. The trained model achieved an accuracy of 87.7 % in predicting streetlights in the test dataset. Utilizing this model, a streetlight shapefile has been developed, with post-processing in ArcGIS Pro to refine accuracy based on aerial imagery. First, the streetlight object detection results were filtered to remove false positive and false negative cases, and GPS was extracted for the images in which the streetlight was detected. These data were then visualized in ArcGIS Pro, where the streetlight locations were manually adjusted based on the streetlight shadows in the Gwinnett County high-resolution aerial imagery. The accuracy of the obtained and actual GPS coordinates was tested using the field high-precision GPS measuring unit for a sample of streetlights. Additionally, nighttime wheelchair video data was collected to estimate scene luminance, and the results were analyzed to develop a methodology for estimating the street lighting coverage on the sidewalks. Using the street light location dataset, the percentage coverage on the sidewalk links within the study area of the thesis has been estimated. To demonstrate the potential variation in routes obtained when streetlight impedance is added, sidewalk links were classified into five different classes and assigned exaggerated multiplicative time-impedance values ranging from 1.0 to 3.0, based on the percentage coverage. Using these initial impedance values, SidewalkSim was used to perform the shortest path routing analysis for two scenarios (Scenario 1: time only impedance, where only the distance was considered as the factor pedestrian route, and Scenario 2: time + attribute impedance, where street lighting impedance has been incorporated). The routes obtained for these two scenarios were analyzed using a percentage detour metric and it was observed that for the 20 trips analyzed between 5 origins and 4 destinations, the average percentage detour in the study area was 4.48 %.

Degree

thesis:*
Level thesis:degree_level
Masters
Department dc:contributor.department
Civil and Environmental Engineering
Grantor dc:publisher
Georgia Institute of Technology
Year dc:date.issued
2024

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Nerella, Thanmai
Advisor dc:contributor.advisor
  • Guensler, Randall L.
Committee members dc:contributor.committeemember
  • Hunter, Michael P.
  • Guin, Angshuman

Subjects

dc:subject × 6

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/1853/75328

Chain of custody

source
Harvested from
Georgia Tech
Base URL
repository.gatech.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Nerella, Thanmai. Impact of Street Lighting on Pedestrian Path Routing - Impedance Factor Implementation. Masters thesis, Georgia Institute of Technology, 2024. https://hdl.handle.net/1853/75328