Learning of Long-Horizon Sparse-Reward Robotic Manipulator Tasks with Base Controllers

---

This is the official implementation of our TNNLS paper:

"Learning of Long-Horizon Sparse-Reward Robotic Manipulator Tasks with Base Controllers"

Guangming Wang, Minjian Xin, Wenhua Wu, Zhe Liu, and Hesheng Wang

Prerequisites

---

docker: https://hub.docker.com/layers/wenhua231/vsdrl/latest/images/sha256-d09a2e0e1f53417b6206a4a4142b37cbdbd769f59ed6ef462fae6f077f40e3ad?context=repo

• Python 3.6.9
• PyTorch 1.10.1
• CUDA 10.2
• pybullet

Usage

---

Train

Train the complete algorithm with state-input. -t is the training task. -l is the saving log.

python learn.py -q -b -c -t 1 -l 1

Train the complete algorithm with image-input.

python learn.py -q -b -c -i

Train with ensemble Base Controllers.

python learn.py -q -b -c -e

test

Test the model in log -l of task -t.

python test.py -t 1 -l 1 

plot

plot the learning curve for a set of training.

python plot.py

Citation

---

If you find our work useful in your research, please cite:

@article{wang2022learning,
  title={Learning of Long-Horizon Sparse-Reward Robotic Manipulator Tasks With Base Controllers},
  author={Wang, Guangming and Xin, Minjian and Wu, Wenhua and Liu, Zhe and Wang, Hesheng},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  year={2022},
  publisher={IEEE}
}