A Fuzzy Clustering Approach to Identify Pedestrians' Traffic Behavior Patterns

Area of Science:

• Traffic Safety Research
• Behavioral Science
• Urban Planning

Background:

• Pedestrian traffic behavior pattern recognition improves management efficiency and planning.
• Understanding these patterns is crucial for targeted interventions and safety measures.
• This study evaluates pedestrian behavior patterns and associated risk factors.

Purpose of the Study:

• To evaluate pedestrian traffic behavior patterns using a fuzzy clustering algorithm.
• To identify factors associated with higher-risk pedestrian traffic behavior.
• To inform safety measures and pedestrian training initiatives.

Main Methods:

• Utilized fuzzy c-means (FCM) clustering on data from 600 pedestrians in Urmia, Iran.
• Employed the Pedestrian Behavior Questionnaire (PBQ) with 5 domains.
• Applied multiple logistic regression to determine risk factors for traffic behaviors.

Main Results:

• Identified two distinct clusters: lower-risk and higher-risk behaviors.
• The majority of pedestrians (64.33%) exhibited lower-risk behaviors.
• Higher-risk behaviors were significantly associated with younger age (≤33), lower education (≤6 years), male gender, unmarried status, and public transportation use.

Conclusions:

• Fuzzy c-means successfully identified distinct lower-risk and higher-risk pedestrian behavior patterns in Urmia.
• Findings provide valuable insights for policymakers to develop targeted safety interventions.
• The study highlights demographic and behavioral factors influencing pedestrian risk, aiding in focused training programs.