Gaze Behavior and Eye-Hand Coordination in Telemanipulation

For effective and natural robotic assistance, it is essential that robotic systems anticipate human intentions in a timely and reliable manner. Among available behavioral signals, gaze is particularly promising, as eye movements during everyday manipulation are tightly linked to motor actions and typically precede them by hundreds of milliseconds. This predictive nature of gaze has motivated its use in intention-aware robotic systems. However, it remains unclear whether the natural anticipatory and task-driven behavior of gaze is preserved when actions are executed through a robot rather than through one’s own body. Here we address this gap by investigating gaze behavior and eye-hand coordination as participants perform goal-directed telemanipulation tasks. Our findings show that gaze allocation during telemanipulation remains aligned with task goals, indicating that its predictive role is preserved also during the execution of robot-mediated manipulative actions. At the same time, teleoperation induces systematic changes in visual behavior, with increased reliance on monitoring the robotic end-effector and manipulated objects, likely reflecting compensation for altered somatosensory feedback. From a robotics perspective, these findings establish gaze as a reliable cue for human intention prediction during robot-assisted manipulation, highlighting changes induced by teleoperation that should be considered when designing intention-aware control strategies for shared autonomy.