08 Fakultät Mathematik und Physik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/9
Browse
1 results
Search Results
Item Open Access Simulation of large scale pedestrian flow(2015) Dridi, Mohamed H.; Wunner, Günter (Prof. Dr.)Pedestrian simulation is a challenging and fruitful application area for particle simulation, especially in places where many people are gathered (e.g. the Hajj, sports and concert events). Traffic and transportation domains take advantage of this simulation as well. Here the design and implementation involves interesting issues and particle-based modelling allows for the reproduction of pedestrian behaviour to a level of detail beyond pure collision-free locomotion. In this dissertation we will present a prototypical study: a simulation of pedestrian movement throughout the entire Haram area in the city of Mecca, Saudi Arabia, during the busiest Hajj hours. The objectives of this study are twofold: firstly to gather data for accurate predictions for designers and decision makers, and secondly to show, verify and validate that through the use of simulation tools it is possible to reproduce realistic scenarios. Typical questions in pedestrian motion planning and design are: How many minutes would be needed to evacuate a football stadium or a pop concert arena? Where is the best place for pedestrian information or a poster wall to be placed? How many exit signs should be distributed over the area and where should they be placed? How many minutes does a pedestrian need to go from point A to point B in the overcrowded places? Are transfer times still realistic when the pedestrians are not familiar with the layout? For better planning in this area a better understanding of human behaviour is required and one way to achieve this is to improve the tools available for pedestrian planning (such as pedestrian micro simulation). In particular, particle-based simulation has become an attractive paradigm for modelling and simulating pedestrian decision-making and behaviour because it supports a one-to-one correspondence between the subject of observation and the simulated particle. Since pedestrian movement is very complex, a computing method has to be developed to simulate the crowded movement of pedestrians on a virtual area. The method is based on the data and relations published by Predtechenski and Milinski [1], resulting from several thousand observations. In this investigation we have developed two-dimensional space crowd dynamic models to allow for a simulation of high density crowds. This was done by modifying and enhancing various features on existing microscopic models, respecting the properties of the individuals like social force and Cellular Automata models. We have also studied all existing models and their positive and negative properties to describe a real crowd behaviour. Our target was to describe a comprehensive approach to model a large scale of pedestrian scenarios (like pilgrim mobility in Hajj) through integrating a number of models from different research domains. In this dissertation a model based on discrete theory is developed and classical mechanics such as Newton’s laws of motion are used. The sum of all the forces exerted by the particle or acting on this particle is called f. This force requires intensive modelling efforts that will be considered and implemented as force-vectors. In this work we will discuss the main forces that can be identified as acting on a pedestrian. This model was developed from image data (so called empirical data) taken from a video camera or data obtained using human observers. We consider the individuals as self-driven particles interacting through social and physical forces, which is one approach that has been used to describe crowd dynamics. The behaviour of the pedestrians can be observed from video tapes and from the given task cycle. Since not all behaviour of pedestrians has a major global effect on the overall capacity, the main affecting behaviour has to be extracted and used as input for the computer model.