Skip to content

Differentiable Social Force simulation with universal interaction potentials.


Notifications You must be signed in to change notification settings


Repository files navigation

Executable Book documentation.
Deep Social Force (arXiv:2109.12081).

Deep Social Force

Deep Social Force
Sven Kreiss, 2021.

The Social Force model introduced by Helbing and Molnar in 1995 is a cornerstone of pedestrian simulation. This paper introduces a differentiable simulation of the Social Force model where the assumptions on the shapes of interaction potentials are relaxed with the use of universal function approximators in the form of neural networks. Classical force-based pedestrian simulations suffer from unnatural locking behavior on head-on collision paths. In addition, they cannot model the bias of pedestrians to avoid each other on the right or left depending on the geographic region. My experiments with more general interaction potentials show that potentials with a sharp tip in the front avoid locking. In addition, asymmetric interaction potentials lead to a left or right bias when pedestrians avoid each other.

Install and Run

# install from PyPI
pip install 'socialforce[dev,plot]'

# or install from source
pip install -e '.[dev,plot]'

# run linting and tests
pylint socialforce
pycodestyle socialforce
pytest tests/*.py

Ped-Ped-Space Scenarios

Emergent lane forming behavior with 30 and 60 pedestrians:

Download TrajNet++ Data

The Executable Book requires some real-world data for the TrajNet++ section. This is how to download and unzip it to the right folder:

wget -q
mkdir data-trajnet
unzip -d data-trajnet