No Cover Image

Conference Paper/Proceeding/Abstract 276 views 19 downloads

Multi-contextual Analysis for Physiological Behaviour for Estimating Trust in Human-Robot Interaction

ABDULLAH ALZAHRANI, Muneeb Ahmad Orcid Logo

Human-Computer Interaction – INTERACT 2025: 20th IFIP TC 13 International Conference, Belo Horizonte, Brazil, September 8–12, 2025, Proceedings, Part III, Volume: 16110, Pages: 224 - 245

Swansea University Authors: ABDULLAH ALZAHRANI, Muneeb Ahmad Orcid Logo

  • 69923.pdf

    PDF | Accepted Manuscript

    Author accepted manuscript document released under the terms of a Creative Commons CC-BY licence using the Swansea University Research Publications Policy (rights retention).

    Download (1.2MB)

Check full text

DOI (Published version): 10.1007/978-3-032-05005-2_12

Abstract

Existing work on estimating user trust in robotic systems has primarily utilised datasets that monitored variations in physiological behaviours (PBs), evolving from one context of interaction. Consequently,in this paper, we created two datasets from two different human-robot interaction (HRI) contex...

Full description

Published in: Human-Computer Interaction – INTERACT 2025: 20th IFIP TC 13 International Conference, Belo Horizonte, Brazil, September 8–12, 2025, Proceedings, Part III
ISBN: 9783032050045 9783032050052
ISSN: 0302-9743 1611-3349
Published: Cham Springer 2026
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa69923
Abstract: Existing work on estimating user trust in robotic systems has primarily utilised datasets that monitored variations in physiological behaviours (PBs), evolving from one context of interaction. Consequently,in this paper, we created two datasets from two different human-robot interaction (HRI) contexts, namely competitive and collaborative, to explore trust dynamics comprehensively. The datasets consisted of participants’ electrodermal activity (EDA), blood volume pulse (BVP), heart rate (HR), skin temperature (SKT), blinking rate (BR), and blinking duration (BD) across multiple sessions of collaborative HRI during trust and distrust states. We investigated the differences in PBs between trustand distrust states and explored the potential of incremental transfer learning methods in predicting trust levels during HRI using the two datasets. The findings showed significant differences in HR between trust and distrust groups. It further showed that the Decision Tree classifier achieved the best accuracy of 89% in classifying trust, outperforming the previous work, while HR, BVP, and BR were the important features. Overall, the findings indicate the potential for applying incremental transfer learning to real-time datasets collected from different HRI contexts to estimate trust during HRI.
Item Description: Lecture Notes in Computer Science (LNCS, volume 16110)
Keywords: Trust; Measurement; Physiological Behaviour; Human-Robot Interaction; Real-time
College: Faculty of Science and Engineering
Start Page: 224
End Page: 245

Similar Items