Studies of AR-guided tasks are vital for developers to adopt effective techniques that boost productivity and reduce errors. However, many studies use simple tasks with non-professionals in lab settings, which do not reflect real industrial environments or the actual workforce.
The University of Cambridge Cyber-Human Lab (CHL), led by Dr. Thomas Bohné, is addressing this by measuring how research settings impact participant behavior and effectiveness, aiming for more reliable user study outcomes.
One CHL project compared AR-guided manual assembly in lab and real workplace environments. The team developed AR systems for manual assembly, increasing task complexity, and compared results from participants—recruited from MakeUK’s apprenticeship program and representing the target industrial workforce—in both settings.
Another CHL project examined how computer vision-based error detection affects different user groups during manual assembly. The study looked at the balance between too many alerts (causing fatigue) and too few (leading to mistakes), focusing on how prediction errors in AR assembly guides influenced task performance and user perception. Three groups were studied: novices, manufacturing experts, and AR/VR experts, to see how experience levels affected their responses.
Webinar attendees will learn best practices for AR user studies, how user perception compares to actual performance, other CHL research achievements, and future research directions for AR in the workplace. The webinar is designed for enterprise AR project managers, architects, developers, and platform providers.