Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Influence of Gripper Design on Human Demonstration Quality for Robot Learning

Gina L. Georgadarellis, Natalija Beslic, Seonhun Lee, Frank C. Sup, Meghan E. Huber

Abstract

Opening sterile medical packaging is routine for healthcare workers but remains challenging for robots. Learning from demonstration enables robots to acquire manipulation skills directly from humans, and handheld gripper tools such as the Universal Manipulation Interface (UMI) offer a pathway for efficient data collection. However, the effectiveness of these tools depends heavily on their usability. We evaluated UMI in demonstrating a bandage opening task, a common manipulation task in hospital settings, by testing three conditions: distributed load grippers, concentrated load grippers, and bare hands. Eight participants performed timed trials, with task performance assessed by success rate, completion time, and damage, alongside perceived workload using the NASA-TLX questionnaire. Concentrated load grippers improved performance relative to distributed load grippers but remained substantially slower and less effective than hands. These results underscore the importance of ergonomic and mechanical refinements in handheld grippers to reduce user burden and improve demonstration quality, especially for applications in healthcare robotics.

Influence of Gripper Design on Human Demonstration Quality for Robot Learning

Abstract

Opening sterile medical packaging is routine for healthcare workers but remains challenging for robots. Learning from demonstration enables robots to acquire manipulation skills directly from humans, and handheld gripper tools such as the Universal Manipulation Interface (UMI) offer a pathway for efficient data collection. However, the effectiveness of these tools depends heavily on their usability. We evaluated UMI in demonstrating a bandage opening task, a common manipulation task in hospital settings, by testing three conditions: distributed load grippers, concentrated load grippers, and bare hands. Eight participants performed timed trials, with task performance assessed by success rate, completion time, and damage, alongside perceived workload using the NASA-TLX questionnaire. Concentrated load grippers improved performance relative to distributed load grippers but remained substantially slower and less effective than hands. These results underscore the importance of ergonomic and mechanical refinements in handheld grippers to reduce user burden and improve demonstration quality, especially for applications in healthcare robotics.
Paper Structure (16 sections, 4 figures, 1 table)

This paper contains 16 sections, 4 figures, 1 table.

Table of Contents

  1. INTRODUCTION
  2. Overview of UMI and Its Design Challenges
  3. Performance and Usability Limitations
  4. Design Modifications UMI for Bandage Opening
  5. EXPERIMENTAL METHODS
  6. Participants
  7. Experimental Design
  8. Task Performance and Perceived Workload Metrics
  9. Statistical Analyses
  10. RESULTS
  11. Bandages Opened
  12. Bandages Damaged
  13. Time to Open Bandage
  14. Perceived Workload
  15. DISCUSSION
  16. ...and 1 more sections

Figures (4)

  • Figure 1: A: Gripper Conditions. Participants bimanually opened bandages with three different gripper conditions. B: Experimental Protocol. In each trial, participants opened a bandage and removed it from its wrapper.
  • Figure 2: Redesigned finger mount of the UMI handheld gripper. The modification allows the fingers to make direct contact when fully closed, enabling effective grasping.
  • Figure 3: Task Performance Results. Bar graphs of task performance metrics (A: Bandages Opened, B: Bandages Damaged, and C: Time to Open Bandage) averaged across participants in each gripper condition (Hands, H; Concentrated Load Grippers, C; Distributed Load Grippers, D). D: Time to Open Bandage over trials averaged across participants in each gripper condition. Error bars and shaded regions represent the 95% confidence intervals. *, **, ***, and **** indicate that the Bonferroni-corrected planned comparison between gripper conditions was statistically significant with $p\rm{_{adj}< 0.05, 0.01,0.005}$, and $\rm{0.001}$, respectively.
  • Figure 4: Perceived Workload Results. Bar graphs of task performance metrics (A: NASA-TLX overall raw score and B: NASA-TLX subscale scores averaged across participants in each gripper condition (Hands, H; Concentrated Load Grippers, C; Distributed Load Grippers, D). Error bars represent the 95% confidence intervals. *, **, ***, and **** indicate that the Bonferroni-corrected planned comparison between gripper conditions was statistically significant with $p\rm{_{adj}< 0.05, 0.01,0.005}$, and $\rm{0.001}$, respectively.