CASA Centre for Advanced Spatial Analysis 1 10 March, 2005 New Approaches to Urban New Approaches to Urban Modelling Modelling : : Agents, Cells, Representations and Agents, Cells, Representations and Visualizations Visualizations Kay Kitazawa Kay Kitazawa University College London University College London [email protected]http://www.casa.ucl.ac.uk/ CASA Centre for Advanced Spatial Analysis 2 Contents Contents Research at CASA Results Pedestrian behaviour modelling An Integrated Simulation Model of Pedestrian Movements -an application to retail behaviour- PDF processed with CutePDF evaluation edition www.CutePDF.com
30
Embed
Kay Kitazawa - STRC · Kay Kitazawa University College ... (TfL report) 9 CASA Centre for Advanced Spatial Analysis 17 ... Buying motives YES Shop explorer Shop-till-you-drop ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
CASA Centre for Advanced Spatial Analysis 1
10 March, 2005
New Approaches to Urban New Approaches to Urban ModellingModelling: : Agents, Cells, Representations and Agents, Cells, Representations and VisualizationsVisualizations
Kay KitazawaKay KitazawaUniversity College LondonUniversity College London
Centre for Spatial Information Science, University of Tokyo
Bath city
Digital CityDigital City
CASA Centre for Advanced Spatial Analysis 20
Patterns of users’ routes/activitiesNecessary Information - contexts
Needs for Pedestrian behavior model
Provide appropriate information according to user’s location / needs
How to avoidtraffic jam?
Where are mypals?
Routes for wheel chairuser?
Digital city
Positioningtechnology
trajectory
Background: Background: LocationLocation--based servicebased service
11
CASA Centre for Advanced Spatial Analysis 21
• There are several needs to develop pedestrian behaviour models
• Key issuesUnderstand and explain real pedestrian’s
movement
Represent dynamic interaction process betweenpedestrians and their environment
( esp. Information which people obtain )
Requirements of pedestrian behavior modelsRequirements of pedestrian behavior models
CASA Centre for Advanced Spatial Analysis 22
Review on cReview on current pedestrian behavior modelsurrent pedestrian behavior models
Micro scale behaviour (e.g. obstacle avoidance)
Crowd dynamics
Transport model
Stochastic model
Network analysis and OD/route estimation
Probability of state-to-state transition
12
CASA Centre for Advanced Spatial Analysis 23
↑Estimation of the next steps of other pedestrians
← Collision avoidance bahaviour
Current position (xi, yi)Velocity (ui, vi)Radius riNormal walking speed ViDestination (pxi, pyi) (qxi, qyi)speed ratio kiPersonal space ratio ciInformation space (di, di
t )
(Kai Bolay)
Crowd dynamicsCrowd dynamics
CASA Centre for Advanced Spatial Analysis 24
Transport modelTransport model Area: S1, S2…SnTrips between Si to Sj : yijDistance between Si to Sj : dij
Destination
Origin Shortest path between OD
( weights associated with each link can bedistance, costs, condition of the road, etc)
•Influence of other areas?•Which area generates more trips than others?•Why?
Gravity modelαi potential as originβj potential as destination
Most evacuation models adopt this concept
Crowd dynamics Ltd
13
CASA Centre for Advanced Spatial Analysis 25
Logit model ---
Consumer: C1, C2,….Cn
Shop: S1, S2,….Sn
Attribute k of shop Sj: Ajk
Probability of Ci choosing Sj: pij
Distance between Ci and Sj: dij
parameter estimation bymaximum-likelihood method
calculate probability of discrete choice
CASA Centre for Advanced Spatial Analysis 26
Stochastic modelStochastic model
Home
A
B
Marcov chain model
Only the last state determines what will happen next
・Number of people who visit each placevia another ( Trip n : n>1 )
Probability of visiting from one place to another
The observed number of people at their first destination
Probability of being the last destination
total
Home (OD)Place
(node)
home
Trip 0
3
1
Trip 1Trip 2
14
CASA Centre for Advanced Spatial Analysis 27
• Well represent micro-scale physical response
• Dynamic
Crowd dynamics
Transport model
Stochastic model
advantage disadvantage
Not take it into account:• where they are going to and why• pre-fixed route = static model• geographical attributes
•Suitable for description ofselection behavior
Several things can’t be represented:• interaction between others/environment•cognitive process of pedestrian
•Useful for being briefed onhow people move around
•Capable of representing changeability of movements
•Inadequate to small scale movement•Not explain why they choose certain place
New pedestrian behaviour models are needed
Understand and explain real pedestrian’s movementRepresent dynamic interaction process
betweenpedestrians and their environment
Requirements of pedestrian behavior modelsRequirements of pedestrian behavior models
CASA Centre for Advanced Spatial Analysis 28
Requirements of pedestrian behaviour models
Methodologies
Background
Framework of the model
Research objective & Research Design
project update
Review on current models
15
CASA Centre for Advanced Spatial Analysis 29
Research Aim and ObjectivesResearch Aim and ObjectivesTo develop a new pedestrian behavior model
be capable of explaining real pedestrian’s movement
represents dynamic interaction between pedestrians and their environment
Easy-to-understand interface
be validated through comparison between actual trajectories
Every factors should be determined based on observed dataIt can deal with more complex behavior (e.g. shopping )
visualization, To make the model more transferable
To deal with not only pre-determined route-choicebut also people’s cognitive process or other changeable events
It should be different from playing with beautiful animation
CASA Centre for Advanced Spatial Analysis 30
EXODUS
Look different but follow same behaviour rules
16
CASA Centre for Advanced Spatial Analysis 31
Behavior model(simulation)
+Visualization
CASA Centre for Advanced Spatial Analysis 32
Research Aim and ObjectivesResearch Aim and Objectives