Do It Fast, Forget It Fast: How Timing and Limb Visualizations Affect First-Person Augmented Reality Instructions

Paper Title

Do It Fast, Forget It Fast: How Timing and Limb Visualizations Affect First-Person Augmented Reality Instructions

Publication Info

• Topic area: Design and evaluation of first-person augmented reality (AR) instructional systems for procedural tasks.
• Keywords: Augmented reality, first-person instructions, embodiment, memory retention, imitation timing, limb visualization, user experience, task performance, XR design, procedural learning.

Background and Problem

• Problem / challenge: Existing research on AR instructions has demonstrated benefits for procedural tasks but lacks systematic exploration of how timing (parallel vs. sequential) and limb visualizations (hand vs. arm, opaque vs. transparent) affect performance, memory retention, and user experience.
• Significance: Understanding these factors is critical for designing AR systems that optimize learning outcomes, efficiency, and user comfort in real-world applications like assembly, repair, and maintenance tasks.
• Motivation and related work: Prior studies show that first-person AR perspectives improve task performance and embodiment compared to third-person views. However, the effects of asynchronous instruction timing and body visualization on memory, embodiment, and comfort remain unclear. This paper addresses these gaps by systematically investigating these variables.

Solution

• Proposed approach: A controlled user study to evaluate the effects of imitation timing (parallel vs. sequential), limb visualization (hand vs. arm), and limb visibility (opaque vs. transparent) on task performance, memory retention, and user experience in first-person AR instructions.
• Novelty:
  1. Systematic investigation of timing, limb representation, and visibility in asynchronous first-person AR instructions.
  2. Development of a custom AR application and physical machine interface for controlled experimentation.
  3. Empirical insights into the trade-offs between speed, retention, embodiment, and user comfort.
  4. Design guidelines for optimizing AR instructional systems based on study findings.
• Procedure and key techniques:
  • Participants (n = 40) performed a two-handed manual task using AR instructions under 8 conditions (2 × 2 × 2 factorial design: timing × limb visualization × visibility).
  • Tasks involved imitating step-by-step instructions and repeating them from memory.
  • Data collected included task performance (accuracy, completion time), retention (value, order, hand usage), and user experience (mental load, embodiment, social factors, comfort).
  • Statistical analysis used Linear and Generalized Mixed Effects Models to identify significant effects.

Results

• Concrete findings:
  • Timing: Parallel timing resulted in faster overall task completion and higher embodiment, while sequential timing improved memory retention (values, order, hand usage) and subjective comfort.
  • Limb visualization: Hand-only representations enhanced order retention and comfort, though many participants preferred full-arm visualizations for realism and guidance.
  • Visibility: No significant effects on performance, retention, or user experience were observed for opaque vs. transparent limbs.
• Advantage over baselines:
  • Sequential timing outperformed parallel timing in memory retention and comfort.
  • Hand-only visualizations improved retention compared to full-arm representations.
• Experiments / evaluation:
  • Conducted with 40 participants (22 women, 18 men; mean age = 27.95).
  • Task: Two-handed manipulation of a machine interface with 8 conditions (timing × limb × visibility).
  • Metrics: Task accuracy, completion time, retention (values, order, hand usage), mental load, embodiment, social factors, and comfort.
• Limitations and future work:
  • Study focused on a single procedural task in a controlled lab setting, limiting external validity.
  • Predefined limb shapes and skin tones may not fully capture individual embodiment preferences.
  • Future work should explore diverse tasks, naturalistic settings, personalized avatar generation, and varying transparency levels.

Summary

This study systematically investigated how imitation timing, limb visualization, and visibility affect performance, memory retention, and user experience in first-person AR instructions. Parallel timing increased speed and embodiment, while sequential timing improved retention and comfort. Hand-only visualizations supported better order retention, though full-arm representations were preferred by some participants. Visibility had no significant impact. These findings provide actionable design guidelines for balancing speed, retention, and user comfort in AR instructional systems, paving the way for more effective and user-friendly AR learning tools.