Skip to content
/ htpm Public

Human trajectory prediction resources from an AI Research Lab at the University of Maine

Notifications You must be signed in to change notification settings

ssocolow/htpm

Repository files navigation

htpm

This is a project to develop a human trajectory prediction model/method. Inspired by https://github.com/xuehaouwa/SS-LSTM (shtp means standard human trajectory prediction)

We would like to thank the creators of the ETH dataset and the creators of the UCY dataset.

Quick Start

Here is a google collab file to explore: https://colab.research.google.com/drive/1HQgejgDW-bQ4lh8iklCOmB7Q8fzEEOVw?usp=sharing

Or you can run it locally by running these commands:

Download the files
git clone https://github.com/ssocolow/htpm.git

Go into the directory
cd htpm

Make the setup script executable (you may not have to do this step, do if the next step returns a permissions error)
chmod +x setup.sh

Run the setup script
./setup.sh

Train the model
python3 ADI_model.py

Background

These papers lead up to this work https://docs.google.com/document/d/1kWWwZU0jlQkvf-aBQ3mmEMIbwEpNAbTCfJpj0H_DWvU/edit?usp=sharing

"SS-LSTM: A Hierarchical LSTM Model for Pedestrian Trajectory Prediction" by Hao Xue, Du Q. Huynh, and Mark Reynolds: https://www.researchgate.net/profile/Du_Huynh/publication/2269555_Self-Calibrating_a_Stereo_Head_An_Error_Analysis_in_the_Neighbourhood_of_Degenerate_Configurations/links/5c03ccb0a6fdcc1b8d502965/Self-Calibrating-a-Stereo-Head-An-Error-Analysis-in-the-Neighbourhood-of-Degenerate-Configurations.pdf

Disclaimer

The processed data may be wrong

About

Human trajectory prediction resources from an AI Research Lab at the University of Maine

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published