Advancements in Pedestrian Simulation Modelling Insight into the Needs, Data Collection, Calibration and Future of Pedestrian Simulation Models - Planning for the Pedestrian Needs for the 2010 FIFA World Cup
S Ahuja, Capita Symonds, UK
Recent advances in pedestrian simulation modelling are given. Need for disaggregate pedestrian data collection to capture interactions between pedestrians and other modes is explained. We base our findings from recent studies for 2010 FIFA world cup
Pedestrian Modelling is a field gaining increasing attention from transport modellers. In past, transport planner and modellers have concentrated on modelling vehicular modes alone. In most transport models pedestrian needs have been neglected. There is a realisation in transport planning profession, that without adequate importance being given to pedestrians, no model can truly depict the reality. Hence, vehicular simulation models offer solutions that are only partial in nature.
Over the last few years, many advancements have been made in the field of pedestrian simulation modelling. Most pedestrian models try to address how pedestrians evacuate buildings or ships in case of emergency and fire. They also look at capacity issues at bottlenecks such as rail interchanges. Recent advances in computational technologies have lead to development of simulation models ranging from macroscopic models (such as Pedflow, Space Syntax methods, Pedroute), mesoscopic models using cellular automata (such as STEPS, PedGo) to truly disaggregate microscopic models using agent based modelling (such as Legion, VISSIM and CAST)
The motivation to model pedestrians comes from the fact that for all journeys we are pedestrians. Hence, one cannot ignore the pedestrian needs in transport projects. In addition, there is a shift towards providing greater pedestrian priority and safety at junctions and interchanges and town centres. Thus, one needs design for pedestrians not only in evacuation conditions but also in normal conditions. This leads to the need to understanding pedestrian psychology and modelling pedestrian behaviour accurately, weather they are inside buildings, or outside in open spaces or interacting with traffic infrastructure at traffic signals or at public transport interchanges.
We present our research in pedestrian simulation modelling, mainly concentrated on agents based models (using VISSIM). In this paper, we describe the development of dynamic pedestrian space models, where agents (pedestrians) react to other pedestrians based on the properties of the areas, which they occupy. For example, pedestrians in lift can tolerate closer spaces, compared to queuing conditions, compared to larger space needs in open walking conditions. This model is vital for space designing and is a major advancement over the grid based cellular automata model. We also present development of elevator allocation model, which takes into account movement of pedestrians in multi-story buildings, based on elevator floor allocation logic. This model is vital in determining the need for the number of lifts in multi-storey buildings and optimum waiting times for lift passengers. We present, advances in optimising traffic signals and pedestrian junctions based on realistic pedestrian behaviour such as tendency to jump traffic lights or wait at signals if people have children or if their mobility is impaired. We present our research into optimising timetables for public transport services and the design of pedestrian interchanges based on pedestrian dispersion within interchanges. Finally, we summarise our findings from our recent psychological research into pedestrian route choice in open areas and how people decide the paths taken for their pedestrian journeys.
The paper focuses on the needs for disaggregate pedestrian data collection that can effectively capture interactions between pedestrians, cyclists and mechanised modes of travel. Methods in data capture, such as automated video capture for counting pedestrians, estimating disaggregate walking speeds and understanding their observed their behaviour corresponding to their space needs are presented. In addition, the use of manual counts, pedestrian tracking surveys to capture the OD movements, face-to-face interviews to understand pedestrian? psychological behaviour and reasoning for their behaviour is explained. The paper the data analysis tools and its representation in agent based models.
We present the use of above data in developing and calibrating the pedestrian simulation models. The calibration variables include microscopic variables such as pedestrian density, flows, throughputs shape and movements of bottlenecks and larger macroscopic variables such a comparison of travel times, evacuation times and perception of level of service.
To support the above work, we base our findings from our recent pedestrian studies in South Africa where we redesigning 8 major rail stations and bus interchanges, in Durban and Cape Town in preparation of the FIFA 2010 world cup.
Finally, we look into the future of pedestrian simulation models and the future requirements for pedestrian simulation models, which include correctly representing not only individual but also group behaviour. We present the attempt to integrate pedestrian simulation into highly realistic virtual reality models that can used for correct 3-dimensional visualisation, better architectural design and evaluation of structural strength of facilities. The paper identifies areas for further research and data collection methods for pedestrian simulation and adoption of standards for its proper calibration and validation.
Association for European Transport