Increasing Input Accuracy of Embodied Devices via Electrical Muscle Stimulation

Paper Title

Increasing Input Accuracy of Embodied Devices via Electrical Muscle Stimulation

Publication Info

• Topic area: Enhancing input accuracy in embodied devices using electrical muscle stimulation (EMS).
• Keywords: Embodied devices, electrical muscle stimulation, input accuracy, proprioception, interaction techniques, recall, confirmation, constraints, user study, gestural input.

Background and Problem

• Problem / challenge: Embodied devices, which rely on the user's body for input and output, lack effective techniques to align users' proprioceptive inputs with the interface state, leading to inaccuracies and usability challenges.
• Significance: Improving input accuracy in embodied devices is critical for enabling reliable, eyes-free interaction in mobile or demanding contexts, such as walking or multitasking.
• Motivation and related work: Prior research explored EMS for proprioceptive feedback but did not formalize or evaluate techniques to improve input accuracy. Existing methods often conflict with users' movements or distract from tasks, leaving a gap in usability enhancements for embodied devices.

Solution

• Proposed approach: Three interaction techniques—recall, confirmation, and constraints—are introduced to improve input accuracy in embodied devices by leveraging EMS for proprioceptive feedback.
• Novelty:
  1. Systematic definition and evaluation of EMS-based interaction techniques for embodied devices.
  2. Demonstration of significant improvements in input accuracy and user confidence through a user study.
  3. Application of Nielsen’s usability heuristics (e.g., visibility of system status, error prevention) to proprioceptive interfaces.
• Procedure and key techniques:
  • Recall: Aligns the user's limb with the interface state upon invocation by actuating the limb to the correct position.
  • Confirmation: Provides proprioceptive cues (e.g., detents) to signal valid state transitions during input.
  • Constraints: Physically bounds user inputs to valid ranges, preventing errors and signaling limits.

Results

• Concrete findings:
  • Input accuracy improved by 40.91% (mean error reduced from 34.61° to 20.45°) in the combined condition compared to the baseline.
  • User confidence increased by 53.85% (mean Likert rating rose from 3.25 to 5.00) in the combined condition.
  • Recall alone significantly improved accuracy for the first target after distraction (mean error reduced to 9.82° vs. 19.60° baseline).
• Advantage over baselines:
  • Combined techniques outperformed individual techniques and the baseline in accuracy and confidence.
  • Recall and confirmation were particularly effective in aiding users' mental models and reducing cognitive load.
• Experiments / evaluation:
  • A user study with 12 participants tested five conditions (baseline, recall-only, confirmation-only, constraints-only, combined) using a wrist-based slider.
  • Metrics included absolute and relative input accuracy, confidence ratings, and qualitative feedback.
  • A technical evaluation confirmed the EMS system's accuracy (mean error: 4.83°) and reliability.
• Limitations and future work:
  • Study conducted in a stationary lab setting, limiting generalizability to mobile contexts.
  • Focused on slider-type devices; future work should explore other embodied device types (e.g., buttons, knobs).
  • EMS accuracy constraints (4.83° error) may limit precision in some applications.

Summary

This paper introduces and evaluates three EMS-based interaction techniques—recall, confirmation, and constraints—to improve input accuracy in embodied devices. A user study demonstrated that combining these techniques significantly enhanced accuracy (40.91% improvement) and confidence (53.85% increase) compared to baseline proprioceptive input. While the study focused on a wrist-based slider, the findings suggest broader applicability to other embodied device types and contexts. Future work should explore these techniques in mobile settings and with diverse device designs.