Context-Continuous Preference Learning for Exoskeleton Personalization
A Gaussian-process preference model shares nearby-context feedback to personalize ankle and elbow exoskeleton assistance with less data.
Context-Continuous Preference Learning uses a Gaussian process to share preference observations across nearby operating conditions while retaining context-specific utility estimates for exoskeleton assistance. Simulations improved reconstruction and preference-based Bayesian optimization when preferences varied smoothly, but produced negative transfer when continuity was weak. In retrospective studies of nine healthy adults, five exposures per context raised mean reconstruction correlation from 0.644 to 0.720 for ankle assistance and from 0.476 to 0.526 for elbow assistance versus independent learning. Benefits for online personalization in humans were not established.
- CCPL shares Gaussian-process preference data across nearby operating contexts.
- Ankle reconstruction correlation rose from 0.644 to 0.720 with five exposures.
- Elbow correlation rose from 0.476 to 0.526 versus independent learning.
- Simulation showed negative transfer when preference continuity was weak.
From the original article
1 min read at arxiv.orgPersonalizing exoskeleton assistance across operating conditions is constrained by the time and physical effort required to collect user feedback. We examined whether a user's preference landscape varies smoothly across operating conditions and when this continuity supports learning from limited feedback.
By Sunin Baek · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.