Research License — Safety-RL

Goal directed RL with Safety Constraints for efficient learning when navigating constrained environments.

In this paper, we consider the problem of building learning agents that can efficiently learn to navigate in constrained environments. The main goal is to design agents that can efficiently learn to understand and generalize to different environments using high-dimensional inputs (a 2D map), while following feasible paths that avoid obstacles in obstacle-cluttered environment. We test our proposed method in the recently proposed \textit{Safety Gym} suite that allows testing of safety-constraints during training of learning agents. The provided python code base allows to reproduce the results from the IROS 2020 paper that was published last year.

  •  Ota, K., Sasaki, Y., Jha, D., Yoshiyasu, Y., Kanezaki, A., "Efficient Exploration in Constrained Environments with Goal-Oriented Reference Path", IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), November 2020.
    BibTeX TR2020-141 PDF Software
    • @inproceedings{Ota2020nov,
    • author = {Ota, Kei and Sasaki, Yoko and Jha, Devesh and Yoshiyasu, Yusuke and Kanezaki, Asako},
    • title = {Efficient Exploration in Constrained Environments with Goal-Oriented Reference Path},
    • booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    • year = 2020,
    • month = nov,
    • url = {https://www.merl.com/publications/TR2020-141}
    • }

To download the software, please enter some information about yourself, then review and agree to MERL's research-only licensing terms. (Information is collected for auditing purposes.)

    Software:
    Name:
    Affiliation:
    Street Address:
    City, State, Country, Postal Code:
    E-Mail:
    Confirm E-Mail:
    IP Address: