CrowdMovmentSimulation

Simple real-time agent-based simulation of crowd movement.

Implemented models:

• Discrete preprocessed shortest path agent movement
• Discrete gradient based agent movement

What do you need?

• System supporting OpenGL (Almost any nowadays)
• Python interpreter

Getting Started

1. Clone repository to your machine

   git clone repository_url

2. Install required python packages

   pip install -r requirements.txt

3. You should be ready to go now! run gradient_main.py from project root (Linux)

   ./gradient_main.py

Project Goal

The goal of the project was to showcase and test different methods of crowd simulation. We wanted to implement few different methods, compare one to another and at the end compare them with real-world crowd behavior.

Result

• Now there are two different discrete simulation methods implemented, presenting a different approach to the subject.
• Both methods are really quick.
• Gradient-based model gave much more believable results as far as the agent behavior goes.
• Worth noting is model parameters have a huge impact on simulation results.

Future

The project is in constant development. We plan to:

• Do more testing with different parameters
• Improve the presentation aspects
• Optimize simulations and visualization speed
• Add GUI with tools for easier testing
• Add map creator
• Add different simulation models (We are especially interested in adding a continuous model)

New map generating

If you want to generate your own map do it by using model.environment.environment.direction_map function. Our example can be found in resources.create_map_100x100. (For map greater than 100x100 it takes a while, so don't worry, it works)

Built Using

• PyOpenGL
• NumPy
• glfw

Authors

• Miłosz Blasiak mblasiak
• Piotr Kędra potrkedra
• Mateusz Pitura Piturzasty

License

Acknowledgments

• A Cellular Automaton Model for Crowd Movement and Egress Simulation (Hubert Ludwig Klüpfel)

• A Simple and Realistic Pedestrian Model for Crowd Simulation and Application (Wonho Kang and Youngnam Han )

• Simulation of Pedestrian Crowds in Normal and Evacuation Situations (Dirk Helbing , Illes J. Farkas , Peter Molnar , and Tamas Vicsek )

• Directing Crowd Simulations Using Navigation Fields (Sachin Patil,Jur van den Berg,Sean Curtis,Ming Lin,Dinesh Manocha)