PECAN: Personalizing Robot Behavior through a Learned Canonical Space

This repository provides our implementation of PECAN for the carlo enviroment.

Install

Clone this repository using

git clone https://github.com/VT-Collab/PECAN.git

Testing trained model

We provide the trained model used in our user study (see video here).

Generate driving styles by running.

python get_driving_styles.py

This will show you an example simulation for both tasks (highway and intersection) and ask your for a user id. It then generates 4 practice styles and 4 teaching styles (e.g. [speed, distance] = [95, 10]). It is possible to skip any practice styles by pressing 'escape' on the keyboard.

Run PECAN's interface

python run_interfaceA.py

Training model

The training dataset consist 24 demonstrations. 4 labeled demonstrations corresponding to the extreme styles and 8 unlabeled demonstrations per tasks. The provided model can be re-trained from the following script

python train_model.py