A Human-Centered Safe Robot Reinforcement Learning Framework with Interactive Behaviors

TL;DR

This paper proposes a human-centered Safe Robot Reinforcement Learning framework that incorporates interactive behaviors to improve safety, collaboration, and alignment between humans and robots, addressing key challenges in real-world robotic applications.

Contribution

It introduces a novel framework integrating interactive behaviors into SRRL, highlighting research gaps and proposing directions for robustness, efficiency, transparency, and adaptability improvements.

Findings

Identifies research gaps in safe exploration, value alignment, and collaboration.

Highlights the importance of interactive behaviors like conversational AI in SRRL.

Discusses open challenges for future research in robustness and transparency.

Abstract

Deployment of Reinforcement Learning (RL) algorithms for robotics applications in the real world requires ensuring the safety of the robot and its environment. Safe Robot RL (SRRL) is a crucial step towards achieving human-robot coexistence. In this paper, we envision a human-centered SRRL framework consisting of three stages: safe exploration, safety value alignment, and safe collaboration. We examine the research gaps in these areas and propose to leverage interactive behaviors for SRRL. Interactive behaviors enable bi-directional information transfer between humans and robots, such as conversational robot ChatGPT. We argue that interactive behaviors need further attention from the SRRL community. We discuss four open challenges related to the robustness, efficiency, transparency, and adaptability of SRRL with interactive behaviors.