Use this URL to cite or link to this record in EThOS:
Title: Agent-based modelling of public space activity in real-time
Author: Cheliotis, Kostas
ISNI:       0000 0004 7660 8178
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Understanding how urban space is used by its inhabitants is vital in improving the overall quality of a city's built environment, as it can highlight needs and requirements of everyday life to be addressed in any urban development. Our investigation of urban activity is often approached through spatial models and simulations on the one hand, and urban data on the other. The work presented here explores potential combinations of the two, by coupling urban models with real-time urban data feeds for continuous short-term forecasting of urban activity. This aim is approached through the development of a model of activity in urban public spaces using the agent-based modelling paradigm, calibrated to real-time data input, and applied to the simulation of current activity in public spaces at a fine spatio-temporal scale. Observations about human spatial behaviour are identified in the literature on public spaces and implemented within a 3D modelling framework, thereby extending existing pedestrian and crowd agent-based modelling approaches. Furthermore, a review and evaluation of real-time data feeds pertaining to activity in public spaces is performed, focussing on open and publicly available datasets, and a forecasting model is developed using social media and other datasets as a proxy for current user activity. The resulting real-time model of public space activity is then evaluated through two case studies focussing on two major urban parks in London, UK. The model performs well in capturing park visitor activity in terms of spatial dispersion. Real-time data feeds examined are found to be capable of capturing park visitor activity to some degree; however they are found to be inadequate in supporting a fully fledged, detailed real-time model of public space activity. Finally, potential future trajectories of the approaches are identified in the increasing availability of online 3D mapping data when combined with advances in computational efficiency and data availability, in extending current data visualisation approaches into expansive, fine-scale simulations of real-time urban activity.
Supervisor: Hudson-Smith, A. ; Batty, M. ; Wilson, D. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available