AR Guides
& Task Speed

Context

Three-dimensional interaction is fundamental in modern computing, from CAD and scientific visualization to VR and AR. Applications like architectural planning, industrial prototyping, and medical training increasingly rely on the ability to manipulate virtual objects in 3D environments. As immersive technologies like Augmented Reality (AR) become more widespread, the demand for effective and user-friendly 3D interaction techniques intensifies. However, working with 3D objects presents significant challenges. Efficiency, measured as Task Completion Time (TCT), is crucial, as slow interactions can hinder usability and technology adoption. Ambiguous depth perception and complex control mappings for six degrees of freedom (DOF) further complicate precise positioning.

Research Question & Hypothesis

3D manipulation in AR is still difficult due to imprecise touch and mid-air controls, lack of haptic feedback, and tracking errors. Visual guides like grids and snapping can improve speed, but may reduce accuracy depending on context. The study tests whether guides help users position 3D objects faster in isometric AR simulations.

"How does the presence of visual guides (grids, reference guides) affect the execution speed when positioning 3D objects in AR?"

Methodology

The study used a between-subjects design to evaluate how visual guides affect task completion time (TCT) in a simulated 3D positioning task. Fourteen participants were randomly assigned to either a “With Guides” or “Without Guides” group. The independent variable was the presence of visual guides, and the dependent variable was execution speed measured in seconds.

A custom-built web app simulated a 3D isometric environment where users arranged virtual cubes to match a reference image. Participants used mouse dragging and the Shift key to manipulate objects along three axes. After receiving instructions, they completed the task while their time was automatically tracked. Each session included a usability test and debriefing, lasting around 25 minutes.

Interface Conditions & Prototype Software

The custom-built web application featured two interface conditions: Condition with Guides: The workspace was covered with a visual grid of thin, light grey lines on a white background, designed to aid precise object alignment and positioning. Condition without Guides: The grid and reference guides were completely deactivated. Participants only saw the X, Y, and Z axis lines and the manipulable objects in a basic workspace.

The prototype was a custom-built web application created specifically for this study using standard web technologies (HTML, CSS, and JavaScript). It featured a 3D simulation that mimicked tools like Ikea Room Planner, offering an isometric view to simulate depth on a 2D screen. The application allowed measurement of task completion time and toggling of visual guides. Participants interacted with flat virtual objects that simulated cubes of varying sizes, moving them along the X, Y, and Z axes (simulating height and position on the base plane).

Results & Subjective Feedback

The study involved 14 participants (6 women, 8 men), primarily aged 35-44 (N=7), with high overall technological familiarity. The primary dependent variable was Task Completion Time (TCT) in seconds. Objective Results (TCT): The Mann-Whitney U test revealed no statistically significant difference in TCT between the "With Guides" group (Mean TCT = 82.0s, Median = 75.0s, SD = 33.91s) and the "Without Guides" group (Mean TCT = 101.15s, Median = 62.0s, SD = 68.63s) (U=31.0, p=.456). The data did not support the hypothesis (H1) that guides would significantly increase speed. This finding contrasts with earlier research on more complex snapping techniques.

Subjective Findings: Despite no objective speed difference, a strong subjective preference for guides emerged. The majority of participants (9 out of 14) found it easier to complete the task with guides. Most participants found the task fairly simple (N=6 Neutral, N=5 Easy, N=3 Very Easy) and were confident in their final object placement (N=5 Quite Sure, N=8 Very Sure). This suggests guides may lessen perceived effort or mental strain, resulting in a more favorable user experience, even without a measurable speedup.

Based on statistical analysis, there was no statistically significant difference in task completion time between participants who used visual guides and those who did not (U=31.0, p=.456). Consequently, the study's data did not support the hypothesis (H1) that guides would significantly increase speed. However, subjectively, most participants preferred using the interface with guides, indicating a potential reduction in perceived effort or mental strain.

Recommendations & Future Work

Designers of 3D interfaces or AR simulations should not assume that straightforward visual instructions will always speed up execution, despite subjective preference. It is recommended to conduct specific empirical tests assessing accuracy and speed within the finished application's context, preferably with larger samples and in a real AR environment. Future research should also measure accuracy to identify speed-accuracy trade-offs, examine various guide types (e.g., dynamic or context-adaptive), and investigate factors contributing to subjective preference (e.g., through cognitive load measurement or qualitative interviews).