NOTE: A new version of the Trajectron has been released! Check out Trajectron++!

The Trajectron: Probabilistic Multi-Agent Trajectory Modeling with Dynamic Spatiotemporal Graphs

This repository contains the code for The Trajectron: Probabilistic Multi-Agent Trajectory Modeling with Dynamic Spatiotemporal Graphs by Boris Ivanovic and Marco Pavone.

Installation

First, we'll create a conda environment to hold the dependencies.

conda create --name dynstg python=3.6 -y
source activate dynstg
pip install -r requirements.txt

Then, since this project uses IPython notebooks, we'll install this conda environment as a kernel.

python -m ipykernel install --user --name dynstg --display-name "Python 3.6 (DynSTG)"

Now, you can start a Jupyter session and view/run all the notebooks with

jupyter notebook

When you're done, don't forget to deactivate the conda environment with

source deactivate

Scripts

Run any of these with a -h or --help flag to see all available command arguments.

• code/train.py - Trains a new Trajectron.
• code/test_online.py - Replays a scene from a dataset and performs online inference with a trained Trajectron.
• code/evaluate_alongside_sgan.py - Evaluates the performance of the Trajectron against Social GAN. This script mainly collects evaluation data, which can be visualized with sgan-dataset/Result Analyses.ipynb.
• code/compare_runtimes.py - Evaluates the runtime of the Trajectron against Social GAN. This script mainly collects runtime data, which can be visualized with sgan-dataset/Runtime Analysis.ipynb.
• sgan-dataset/Qualitative Plots.ipynb - Can be used to visualize predictions from the Trajectron alone, or against those from Social GAN.

Datasets

The preprocessed datasets are available in this repository, under data/ folders (i.e. sgan-dataset/data/).

If you want the original ETH or UCY datasets, you can find them here: ETH Dataset and UCY Dataset.