Introduction
Implementation of the paper A Steering Algorithm for Redirected Walking Using Reinforcement Learning
Acknowledgment
This repo is a basic implementation of the paper A Steering Algorithm for Redirected Walking Using Reinforcement Learning. Limited by ability and time, I cannot guarantee that the code is completely consistent with the origin paper, only the basic framework can be provided. This implementation is based on fantastic implementaion PPO by Kostrikov.
Training
- Update your env configuaration like [w, h] of your physical space in srl_core/envs_general.py
- Fine-tune your hyper-parameters in
args/train.txt
- Training a new model
python main.py --config args/train.txt
Evaluation and plot the heatmap
-
python test_heatmap.py --config args/eval.txt --load_epoch $Your_MODEL