ABSTRACT
Crowd disasters can be eradicated if the crowd’s behaviours and movements are understood and can be predicted. Crowd modelling relates to the
development of the dynamical models of the crowd. It is an extremely useful tool to study crowd behaviours and patterns where experiments with
human are too dangerous or impossible. By understanding how people or crowd move through an area or space, better crowd management plans
can be developed to keep people in the crowd safe in time of any danger, whether natural or manmade. This model is also instrumental in, for
example, assessing or revitalizing building designs by architects, designers and safety engineers.
METHOD
Microscopic approach
Social Force Model
Magnetic Model
SESAK™
A CROWD
SIMULATOR
• Describe the agent movement and characteristics
• Path finding in
agent movement
Dijkstra
Algorithm
• Integrated with
AI technique
SIMULATION RESULTS
a. Lane Formation
1. Crowd movement at :
• Intersection
• Bi-directional flow
• Emergency (from 1 exit)
2. Crowd simulator in 3D :
• Different angle view
• Different environment
• Multi-characteristics agent
3.
•
•
•
•
•
Analysis in 2D (Matlab)
Time taken (for emergency)
Flow rate
Self-organized behaviours
Agent speed profile
Path finding
3D Environment
b. Counter Flow
Movement
2D Environment
c. Emergency at 1 Exit
COMMERCIALIZATION POTENTIAL
1.
2.
A tool to minimize congestion and avoid overcrowding
Analysis for transportation or building management services for
security and safety purposes
ASSOC. PROF. IR DR HAZLINA SELAMAT,
CENTRE FOR ARTIFICIAL INTELLIGENCE & ROBOTICS
(CAIRO),
FACULTY OF ELECTRICAL ENGINEERING,
UNIVERSITI TEKNOLOGI MALAYSIA
NURULAQILLA KHAMIS (STUDENT)
CENTRE FOR ARTIFICIAL INTELLIGENCE & ROBOTICS
(CAIRO),
MALAYSIA-JAPAN INT. INST. OF TECH (MJIIT),
UNIVERSITI TEKNOLOGI MALAYSIA