Feasibility Study of Assessing Pedestrian Discomfort Using Behavioral Data

Understanding and assessing pedestrian experiences are essential for creating walkable cities. However, traditional assessment methods, such as surveys, have limitations, highlighting the need for novel approaches. This study proposed a novel approach that involves utilizing pedestrian behavior data to evaluate pedestrian discomfort and explored its feasibility. Pedestrian coordinate data from 40 participants in a real urban environment were collected, and an anomaly detection model was employed to identify abnormal gait patterns. Correlation analysis confirmed that participants’ discomfort was positively correlated with the anomaly point density ratio, suggesting that participants’ discomfort was expressed in behavioral changes. These findings underscore the potential of the proposed approach for assessing discomfort using behavioral data. Furthermore, the study suggested future research directions, such as real-time data acquisition via CCTV systems, to enhance pedestrian discomfort assessment. This study presents a method for assessing and monitoring pedestrian discomfort, which can contribute to creating pedestrian-friendly cities.