Empirical characterisation of agents’ spatial behaviour in pedestrian movement simulation

Route choice behaviour is a key factor in determining pedestrian movement flows throughout the urban space. Agent-based modelling, simulation paradigm that allows modelling individual mechanisms to observe emergence of macro-level patterns, has not employed empirical data regarding route cities or accommodated heterogeneity. The aim this paper present an empirically based Agent-Based Model (ABM) accounts for behavioural heterogeneity strategies, simulate pedestrians cities. We designed questionnaire what degree people employ salient elements (local and global landmarks, regions, barriers) road costs (road distance, cumulative angular change) characterise agent our ABM. hypothesised heterogeneous ABM configuration on construction typologies from would portray more plausible picture than homogeneous configuration, same data, random configuration. city Münster (DE) was used as case study. From sample 301 subjects, we obtained six clusters differed relation role (distant barriers, regions) meaningful local along route. directed agents towards natural streets historical centre. configurations resulted lower volumes roads cars (25% decrease) but higher concentrations Promenade lake (40% increase); knowledge, deem these results plausible. Minor differences were identified between configurations. These findings indicate inclusion does make difference terms patterns. Yet, demonstrated models should be at least average sample-level inform planners about areas prone high pedestrians.