THESIS ABSTRACT

Abstract

Unsignalized intersections in the Kathmandu Valley provide major operational issues for pedestrians due to mixed traffic and a lack of official control mechanisms. This study aims to determine the pedestrian critical gap and evaluate Pedestrian Level of Service (PLOS) at selected crosswalks of unsignalized intersections. The research evaluated average waiting time and variations across demographic factors like gender, group size and carrying luggage while developing log-linear regression-based model to predict waiting time. The methodology involved a video-graphic survey across nine crosswalk legs at three intersections: Pepsicola, Dhapakhel Dobato, and Patan Hospital. The Probability Equilibrium Method (PEM) was used to estimate critical gap, selected for its statistical strength and suitability for mixed traffic compared to other deterministic methods. Total of 2609 observations were analyzed. The results indicated that pedestrian critical gap varied between 2.53s and 4.34s across intersection legs. Average waiting times ranged from 8.53s to 28.05s. Local PLOS boundaries were calibrated at < 5.8s (PLOS A), 5.8–11.7s (PLOS B), 11.7–18.8s (PLOS C), 18.8–27.2s (PLOS D), 27.2–37.7s (PLOS E), and > 37.7s (PLOS F) using K-means clustering method. Under these calibrated benchmarks, intersections at Pepsicola (15.77s) operated at PLOS C, while Dhapakhel Dobato (20.58s) and Patan Hospital (22.19s) both operated at PLOS D. Log-linear multiple regression model was developed to predict waiting time, that yielded an R 2 of 0.515 (Adjusted R 2 = 0.513) with strong statistical significance (F = 276, P < 0.001). The analysis identified critical gap, roadside friction score, roadway width as primary significant predictors of pedestrian delay. Within the log-transformed domain (ln WT), the model demonstrated predictive capability, yielding MAPE of 8.01%. The study concluded that PEM and K-means clustering provide reliable framework to understand pedestrian decision-making and that PLOS at unsignalized crossings need pedestrians crossing facilities improvements, to enhance urban mobility and safety.

Keywords

Pedestrian Critical Gap, PLOS, Probability Equilibrium Method (PEM), K-means Clustering, Unsignalized Intersection, Kathmandu, Waiting Time, Error Metrics.