This repo is for the IROS 2021 paper "Efficient Task Planning for Mobile Manipulation: a Virtual Kinematic Chain Perspective". This repo only contains the PDDL task planning module, the motion planning part is included in this repo.
We have tested this project under Ubuntu 18.04 using Python 3.7. For running this PDDL task planning module, no other dependency is required.
Run the PDDL problem solver
python solve.py <domain-file> <problem-file>
For example, you can use following command to run an example
python solve.py picknplace/domains/domain.pddl picknplace/samples/sample.pddl
The example outputs should be
Generating pick-and-place problems
You can use this scripts to automatically generate a bunch of PDDL pick-and-place problems. More details please refer to the script.
Run the experiment
You can use the following command to re-run the experiment 1
python run_exp1.py <domain-file> <problem-file-dir> <output-dir>
For example,
python run_exp1.py picknplace/domains/domain.pddl picknplace/problems picknplace/results
@inproceedings{jiao2021efficieint,
title={Efficient Task Planning for Mobile Manipulation: a Virtual Kinematic Chain Perspective},
author={Jiao, Ziyuan and Zeyu, Zhang and Wang, Weiqi and Han, David and Zhu, Song-Chun and Zhu, Yixin and Liu, Hangxin},
booktitle={IROS},
year={2021}
}