Crowds Simulation

by BSC / Hugo Perez

Develop crowds' behaviour models based on real data.

Hi,

We are an academic group at Barcelona Supercomputing Center - BSC. We are developing a research project about "Crowds Simulation and Visualization."

The crowd simulation can help in the analysis of different events, for example:

- The emergency evacuation
- The diseases propagation
- Or urban planning

Observe and measure the phenomena that occur in these events is possible, but repeat and experience is not.

This factor added to the increase of population in urban areas. It has become a necessity the use of computational models describing human behaviour.

The use of these systems helps us to take preventive measures, reducing risks and costs.

Currently it is difficult to validate the behaviour models, for different reasons, including:

- Validate against real data involves collecting the information of thousands of people.
- Some scenarios are panic situations that can not be replicated.
- Human behaviour is complex and depends on many different factors; the same person may react differently in similar scenarios.

For these reasons, there is no widely accepted criterion in the validation of simulation models.

However, the emergence of new platforms like OpenPaths gives us the opportunity to collect data from thousands of people around the world in different scenarios.

The aim of the project is based on real data:

- Developing new behaviour models.
- Combine with simulated data for scaling the system.
- Validate and refine our models to make them more realistic.

We want to collect data from thousands of people. The volume of participation will define the city in which we focus.

The collected user data is only used for academic research.

We will publish all findings in peer-reviewed scientific journals. We will also disseminate findings via presentation at conferences.

Please participate! Let us harness the potential of our data for the benefit of all.

The data are provided to ensure your safety anonymously.

It's easy to participate, data are collected automatically when you install the application on your mobile and consumes minimal resources.

The aim of the project is based on real data:

- Developing new behaviour models.
- Combine with simulated data for scaling the system.
- Validate and refine our models to make them more realistic.

We implemented our prototypes on private servers with high safety standards, which only we have access.

The collected user data is only used for academic research.

We will publish all findings in peer-reviewed scientific journals. We will also disseminate findings via presentation at conferences.