Attune's self-annotation tool optimizes robot operator attention

A team of researchers from the University of Maryland, University of Virginia, and University of Wisconsin-Madison has developed Attune, a self-annotation tool designed to optimize human-robot collaboration by analyzing operator attention patterns. Published in arXiv as part of UIST '26 proceedings, Attune addresses a critical challenge in multi-robot supervision: managing operator attention across multiple robotic systems in complex environments.

The tool works by tracking operator eye gaze during robot operation, using AI assistance to automatically identify meaningful gaze shifts and annotate their causes. This process generates a summary of operator attention patterns that can be reviewed to calibrate robot behavior. In a user study, participants used Attune to annotate visual triggers that drew their attention, revealing significant variations in gaze patterns and demonstrating the tool's effectiveness in characterizing operator attention needs for better robot interface design.