Improving instruction conveyance in human-to-human collaboration tasks using Augmented Reality

Abstract

Displaying instructions using Augmented Reality (AR) is a widely explored field, especially in the industry. AR applications have been shown to decrease the time required to complete a given task and the number of errors per employee. However, the potential of collaboration for this scenario has not been exhaustively researched yet. There is already research on task conveyance interfaces for collaborative scenarios by employing head-mounted displays (HMDs). To utilize AR, HMDs are an option but there are also mobile devices like smartphones and tablets, which have the advantage of being cheaper and more familiar to users. We extended the previous work with HMDs by developing a mobile interface for human craft use case where two people work together in collaboration. We developed a mobile application for hand-held displays (HHDs) which receives instructions together with the already developed HMD application. The interface furthermore includes features to support collaboration, such as progress bars visualizing the partners and the users own progress. We followed with a user study, mainly comparing the usability and mental demand of the HMD and HHD user interface. Additionally, we derived some future research topics, making the study partially exploratory. The results showed that the HMD condition outperformed the HHD, by scoring slightly higher in usability. Participants favored HMDs for their hands-free operation and more stable tracking, though downsides such as a limited Field of View (FOV) and high learning curve were noted. The qualitative insights further revealed that the HHD interface was preferred for its familiarity. We suggest clearer collaboration roles and addressing tracking limitations for further studies. Despite the HMD interface's higher performance, participants expressed interest in using mobile systems such as ours for home furniture assembly, suggesting avenues for further exploration in AR interfaces.