Towards understanding of human behaviour in crowded spaces

Human behaviour in the real world is important information for developing human behaviour models and simulations. However, it is difficult to capture ‘real’ human behaviour since each human has unique char-acteristics. As part of the AUNT-SUE (Accessibility and User Needs in Transports - Sustainable Urban Environments) project, this research is aimed at understanding individual human behaviour in crowded spaces based on video observation analysis. The video observation analysis employed a video observation method where a multi-mode transportation system in Malaysia was selected as a case study. The observa-tion focus was at an exit door where considerable variety of human movement and behaviour could be observed. Six hours of video recording was conducted covering weekdays, weekends, peak and off-peak times. Almost 19,000 individual humans were observed and categorised into six different behaviours that were determined from the three major human movements of free, opposite direction and same direction movement.