Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Can you pass that tool?: Implications of Indirect Speech in Physical Human-Robot Collaboration

Yan Zhang, Tharaka Sachintha Ratnayake, Cherie Sew, Jarrod Knibbe, Jorge Goncalves, Wafa Johal

TL;DR

This study investigates indirect speech acts (ISAs) in physical human-robot collaboration using a Wizard-of-Oz design with 36 participants and TIAGo across three tasks. By comparing ISA-capable and non-ISA robots, it systematically measures team fluency, goal alignment, trust, and anthropomorphism, complemented by qualitative interviews. The findings show that robots capable of interpreting ISAs significantly improve all four metrics, though effects are task- and context-dependent, and grounded understanding of ISA usage emerges as a key factor in shared goals and rapport. The work advances HRI by highlighting when and how ISAs can enhance collaboration, urging careful integration of direct and indirect speech and advocating for LLM-enabled, context-aware language handling in cobots to boost performance and user experience.

Abstract

Indirect speech acts (ISAs) are a natural pragmatic feature of human communication, allowing requests to be conveyed implicitly while maintaining subtlety and flexibility. Although advancements in speech recognition have enabled natural language interactions with robots through direct, explicit commands -- roviding clarity in communication -- the rise of large language models presents the potential for robots to interpret ISAs. However, empirical evidence on the effects of ISAs on human-robot collaboration (HRC) remains limited. To address this, we conducted a Wizard-of-Oz study (N=36), engaging a participant and a robot in collaborative physical tasks. Our findings indicate that robots capable of understanding ISAs significantly improve human's perceived robot anthropomorphism, team performance, and trust. However, the effectiveness of ISAs is task- and context-dependent, thus requiring careful use. These results highlight the importance of appropriately integrating direct and indirect requests in HRC to enhance collaborative experiences and task performance.

Can you pass that tool?: Implications of Indirect Speech in Physical Human-Robot Collaboration

TL;DR

This study investigates indirect speech acts (ISAs) in physical human-robot collaboration using a Wizard-of-Oz design with 36 participants and TIAGo across three tasks. By comparing ISA-capable and non-ISA robots, it systematically measures team fluency, goal alignment, trust, and anthropomorphism, complemented by qualitative interviews. The findings show that robots capable of interpreting ISAs significantly improve all four metrics, though effects are task- and context-dependent, and grounded understanding of ISA usage emerges as a key factor in shared goals and rapport. The work advances HRI by highlighting when and how ISAs can enhance collaboration, urging careful integration of direct and indirect speech and advocating for LLM-enabled, context-aware language handling in cobots to boost performance and user experience.

Abstract

Indirect speech acts (ISAs) are a natural pragmatic feature of human communication, allowing requests to be conveyed implicitly while maintaining subtlety and flexibility. Although advancements in speech recognition have enabled natural language interactions with robots through direct, explicit commands -- roviding clarity in communication -- the rise of large language models presents the potential for robots to interpret ISAs. However, empirical evidence on the effects of ISAs on human-robot collaboration (HRC) remains limited. To address this, we conducted a Wizard-of-Oz study (N=36), engaging a participant and a robot in collaborative physical tasks. Our findings indicate that robots capable of understanding ISAs significantly improve human's perceived robot anthropomorphism, team performance, and trust. However, the effectiveness of ISAs is task- and context-dependent, thus requiring careful use. These results highlight the importance of appropriately integrating direct and indirect requests in HRC to enhance collaborative experiences and task performance.

Paper Structure

This paper contains 38 sections, 1 equation, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Speaking to Embodied Agents
  4. Verbal Communication during Human-Robot Collaboration
  5. Indirect Speech Acts in Human-Robot Interaction
  6. Hypotheses
  7. User Study
  8. Experimental Design
  9. Apparatus and Setup
  10. Robot's Speech Understanding
  11. Collaboration Tasks
  12. Participants
  13. Procedure
  14. Data Collection and Analysis
  15. Results
  16. ...and 23 more sections

Figures (2)

  • Figure 1: (a) Experiment setup: The participant, robot, and experimenter 1 were all present in the same room. The participant and robot were seated on opposite sides of a table, with the shared workspace located in the centre. Experimenter 1 (Speech Wizard) sat next to the robot, near the emergency button, and operated the robot's speech-WoZ interface. Experimenter 2 (Motion Wizard) was positioned behind a one-side mirror, allowing for a clear view of the room, and was responsible for teleoperating the robot's arm movements. (b) Experiment procedure: Each participant first completed a pre-questionnaire before being assigned to either the ISA or non-ISA group. The participant then performed three tasks with the robot in a counter-balanced order. Before each task, participants watched a tutorial video and read a task description. After completing each task, they filled out a post-task questionnaire. The experiment concluded with a semi-structured interview.
  • Figure 2: Participant responses on their perceptions of the team fluency (a), goal alignment (b), performance trust (c), and the robot's anthropomorphism (d) under different Speech Modes. (*: This item was originally an inverse item according to hoffman2010effects. To make this figure look consistent, we reversed this item ($current\_score = 8-original\_score$).