A new research tool reads where robot operators look, so designers can shape what draws them.

When one human supervises several robots at once, where their eyes land between camera feeds is part of the design problem. A peer-reviewed paper argues it is something the robots themselves can be designed around.

The paper, Attune: A Self-Annotation Tool for Understanding Robot Operator Attention Profiles (ACM DOI 10.1145/3830398.3830514), introduces a small research tool built for the design stage, before any robots are deployed. The premise: if designers know what pulls an operator's gaze, they can shape a robot's behavior to capture, sustain, or relinquish that attention.

Attune runs on recordings of an operator's eye gaze during a supervision task. It automatically flags when the gaze jumps between feeds, helps the operator label what drew each shift, and produces a summary of the operator's gaze pattern. In a small user study, the team found that operators' gaze patterns vary enough to be treated as profiles rather than noise.

Operator attention, the team argues, is itself a design surface alongside latency, field of view, and robot behavior, not just an HR problem of staffing enough humans. The authors are explicit that the study is pre-deployment and small; Attune is a measurement aid, not a deployment product. A GitHub repository hosts the code.