Guiding gaze, evaluating visual cue designs for augmented reality

Abstract

Visual cueing is an interdisciplinary and complex topic. It has garnered interest for implementation with extended reality (XR). Both augmented reality (AR) and virtual reality (VR), are often employed for visual search tasks. Visual search, a paradigm rooted in cognitive psychology (in particular attention theory), can often benefit from cueing interventions. However, there are several potential pitfalls with using cueing techniques in AR; namely, automation bias, clutter, and cognitive overload. These factors are tied to design and implementation choices, such as modality, representation, dimensionality, reference frame, conveyed information, purpose, markedness, or the task domain. Design factors are subject to both the cognitive factors, as well as, technical specifications of the display technology. To address these factors, this work proposes a within-subject four factor design addressing the question how do different cue designs affect visual search performance? Four cueing conditions are used: no cue (baseline), gaze line, 2D wedge, and 3D arrow. Results support the use of cues for visual search, however the gaze line condition provided for the fastest search time, accuracy, and greatest reduction in head rotation. Additionally, the gaze line cue was preferred by participants and was produced more favorable NASA TLX scores.