Reproduce “Human-Robot Mutual Adaptation in Shared Autonomy” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Stefanos Nikolaidis, Yu Xiang Zhu, David Hsu, Siddhartha Srinivasa
A shared-autonomy framework models the human’s willingness to change strategies as an unknown adaptability parameter and incorporates inferred disagreement into robot planning. In a 51-participant assistive-arm study, mutual adaptation improved task performance over simply following user preferences while preserving trust and perceived collaboration compared with a robot that ignored those preferences.
Key insights
- Disagreement is modeled from the human’s perception of the robot’s strategy, rather than solely from the robot’s internal action or policy.
- Temporary opposition can be beneficial when the user is likely to adapt, but persistent opposition is counterproductive when the user is not.
- Mutual adaptation provides a middle ground between autonomous override and unconditional preference following, linking task optimization with trust-sensitive interaction.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.