Visual Guidance in Dynamic Search: How Augmented Reality Cue Properties Shape Performance

Visual search is critical in real-world tasks, from soldiers identifying hazards to drivers locating people or objects. Augmented reality head-mounted displays (AR-HMDs) can aid search by overlaying visual guidance on the user’s field of view. However, it remains unclear which cue properties are most effective during dynamic search, especially when automated cues are erroneous. This study examined how AR cue properties shape performance in a large, naturalistic 3D environment. Thirty-five participants completed a virtual reality (VR) search-on-the-move task (SOTM) with simulated AR cues that varied in spatial precision, configuration, and reliability. Compared with an unaided search, reliable cues substantially improved performance. Dual cues also improved performance compared with single cues, regardless of whether targets were within or outside the participant’s central line of sight. However, incorrect cues produced clear performance costs. These findings suggest that AR guidance can support dynamic search, but reliability and cue properties should be carefully considered.