Abstract
Indoor wayfinding within expansive, architecturally complex retail environments frequently induces cognitive friction, navigational errors, and heightened spatial anxiety due to the absence of reliable satellite positioning and the unintuitive nature of traditional two-dimensional floor plans. This study systematically evaluates the efficacy of a markerless Augmented Reality (AR) navigation system powered by visual-inertial odometry and visual positioning service (VPS) point-cloud alignment against two conventional wayfinding modalities: interactive 2D mobile digital maps and physical static signage. A between-subjects empirical experiment was conducted with 84 participants tasked with completing multi-destination wayfinding challenges across a 45,000-square-meter multi-level commercial retail complex. Navigational performance was assessed through objective metrics including total completion time, path efficiency ratios, and frequency of hesitation events at topological choice points. Affective and cognitive states were quantified using the State-Trait Anxiety Inventory (STAI-6) and the NASA Task Load Index (NASA-TLX). Quantitative results demonstrate that participants utilizing the markerless AR overlay achieved a 28.4% reduction in overall traversal time and a 42.1% decrease in path deviation compared to the 2D digital map group, alongside an 80.2% reduction in junction hesitation events compared to the physical signage cohort. Furthermore, AR users reported significantly lower post-task state anxiety and overall cognitive workload. These findings suggest that directly superimposing egocentric pathing cues onto the physical environment eliminates the mental rotation overhead associated with allocentric maps, mitigating spatial disorientation and cultivating a more seamless, reassuring indoor navigation experience.
