BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//BritCHI//Schedule//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:RPD-Grounded Simulated Users for Pre-screening AR Training Int
 erface Designs: A Unity ML-Agents Approach
BEGIN:VEVENT
UID:paper-session-2-immersive-and-embodied-experiences-rpd-grounded-simulat
 ed-users-for-pre-screening-ar@britchi.uk
DTSTAMP:20261014T154232Z
DTSTART:20261102T140000Z
DTEND:20261102T141000Z
SUMMARY:RPD-Grounded Simulated Users for Pre-screening AR Training Interfac
 e Designs: A Unity ML-Agents Approach
DESCRIPTION:BritCHI 2026 · Paper Session 2: Immersive and Embodied Experie
 nces\nThilakarathna\, Wong\, Nguyen\nEvaluating augmented reality (AR) tra
 ining interfaces typically requires recruiting domain practitioners\, coor
 dinating controlled studies\, and incurring the logistical and ethical cos
 ts of human-subject testing. This paper presents a pre-screening framework
  that uses simulated users to surface interface design trade-offs before h
 uman trials begin. Agents are built with Unity ML-Agents and grounded in t
 he Recognition-Primed Decision (RPD) model\, with distinct reward structur
 es derived from Cognitive Task Analysis of paramedic diagnostic reasoning 
 encoding novice and expert decision strategies. We evaluate three interfac
 e configurations varying in button density\, spatial layout\, and labellin
 g across both agent types\, and characterize behaviour using behavioural i
 nteraction indicators computed from interaction logs and conceptually orga
 nized using a structure inspired by the NASA Task Load Index (NASA-TLX). R
 esults show that dense dispersed layouts impose the highest temporal and s
 earch cost on novice agents\, while deceptive clustering imposes the highe
 st temporal and effort cost on expert agents. Search Complexity Indicator 
 diverged between cohorts in a pattern consistent with RPD-predicted select
 ive attention: expert agents narrowed search as complexity increased while
  novice agents maintained broad attention throughout. These patterns were 
 consistent across behavioural indicator scores and raw interaction metrics
  including path entropy\, inter-click intervals\, and idle ratios. The fra
 mework requires no human participants\, produces reproducible behavioural 
 traces across conditions\, and is intended as a complement to human-subjec
 t evaluation rather than a replacement. Validation against human self-repo
 rted ratings and objective performance data is identified as the essential
  next step.
LOCATION:Brunel University London\, Kingston Lane\, Uxbridge UB8 3PH
URL:https://www.britchi.uk/programme/schedule/paper-session-2-immersive-and
 -embodied-experiences/
END:VEVENT
END:VCALENDAR