Augmented Reality (AR) with Head-Worn Devices for Building Construction
Co-PIs: Dr. Tanyel Bulbul, Dr. Joe Gabbard
Derek Ho, Chris Wakeley, Irene Soto , Ao Chen
Graduate Research
Identify any workload or situational awareness differences between levels of AR fidelity (conformal, non-conformal) in a building construction environment.
This was a pilot study to determine whether situational awareness and workload was could be augmented while performing construction tasks while wearing different types of AR head-wearable devices (HWD). The pilot study was used to determine
Development of Study
Construction environments are noisy, dusty, and susceptible to weather effects. Typically, safety procedures are missed or overlooked. Current approaches to construction safety are dominated by text-based, checklist-type paper tools. These are mediums which could be addressed via useful and intuitive AR applications for various tasks.
AR technology is readily available and there is potential in its application to improving construction operation. But like most other forms of technology there is concern for worker distraction and perception of peripheral stimuli during use of AR devices.
Levels of AR fidelity: conformal vs. non-conformal
Two different AR displays are used to showcase levels of AR fidelity: conformal and non-conformal. A conformal view is shown on the left in which a 3-D model can be "projected" onto a space within an environment and can be viewed as if it were there in real life. A non-conformal view is shown on the right in which an image remains static and in the user's field of view.
For our study we used ODG R-7 glasses and Vuzix M100 glasses to respectively represent conformal and non-conformal mediums to the user.
ODG R-7 and Vuzix M100 glasses
Identifying measures for situational awareness and workload
To fully understand the task of framing a wall, interviews were performed with experienced construction workers as they detailed the tasks they performed when framing and sheathing walls.
To further understand specific tasks a construction worker would go through, videos of wall framing activity was watched and discussed.
A hierarchical task analysis was then performed to develop situational awareness requirements for the task. This was accomplished through interviews with experts who have performed framing and sheathing tasks in construction.
Previous work have identified multiple ways to measure situational awareness and workload. For this pilot study we are only investigating comparative measures between the different modalities technology (conformal, non-conformal, paper).
Three methods of acquiring the measures will be used for the study:
1. Situation Awareness Global Assesment Technique (SAGAT): freeze-test query technique developed by Mica Endsley that objectively tests the perception, comprehension, and prediction ability of the participant during the task performance.
2. Situation Awareness Rating Technique (SART): post-test questionnaire that subjectively tests the participant's ability to recognize the demand of attentional resources, supply of attentional resources, and understanding of the situation.
3. NASA- Task Load Index (TLX): a post-test tool administered to participants to subjectively determine their workload during task performance in six categories: Mental Demand, Physical Demand, Temporal Demand, Performance, Effort, and Frustration.
Project Design
Development of the questions administered for the SAGAT and SART questionnaires were related to environmental and personnel activity that occurred during the test. Environmental hazards were implemented as "holes" in the platform they were building on as well as "edges" of the platform. Personnel activity included researchers donning a hard hat and hazard vest and pushing a wheelbarrow and transporting wood while the participant was constructing the frame.
A single-factor, randomized, within-group design was used in this study. Independent variables consisted of the blueprint medium, frame to construct, and topic for the subjective SA questionnaire. 6 participants were selected who had construction experience and were affiliated with the Building and Construction Science program. Dependent variables in this study include the subject and objective SA scores determined from the questionnaires, the performance of the framing and sheathing tasks, and tracking information. Each participant completed three tasks, with each task using a different blueprint tool. The different methods used to present the blueprint to the participant were: a non-registered blueprint on a monocular Vuzix M100 device, registered graphics displayed on an ODG R-7 device, and conventional blueprints on a piece of paper. Due to the constraints of the ODG R-7 glasses, participants were required to have 20/20 vision without the use of glasses. To prevent learning effects in building the wall frame from task to task, three similarly complex layouts were designed and each participant had to complete all three designs. To prevent learning effects from the objective questionnaires, the SAGAT questionnaires were broken up into the three different topics: other workers, hazards and piles, and frame and boundary questions. The independent variables were partially counter-balanced due to only having six participants
Visual sample of the visual stimuli blueprint used
Data Collection
Participants were tasked with performing the framing and sheathing of a blueprint provided in the three different modalities (conformal, non-conformal, paper). Objective data was collected (task completion time, accuracy of placement, number of nails used) along with the SA and workload questionnaires mentioned above.
Participants were counterbalanced to avoid any bias in their responses.
SAGAT questionnaires were administered in the midst of the wall framing activity, while SART and the NASA-TLX were administered afterwards.
Graduation, End of Research
Unfortunately after the round of data collection was performed for the pilot study, I was unable to continue working on this project due to my graduation. The good news to come from this was that there was enough evidence to warrant the CO-PIs to continue the study. Relying on my background research on situational awareness and the project study setup, from project design and SA measures, the project is still on-going.
References
[Bulbul, T. "An Analysis of Construction Worker Safety during Building Decommissioning and Deconstruction" Proceedings of the CIB W099 Safety and Health in Construction Conference, August 2011, Washington, DC.]
[1] Liu, S. A. Jenkins, P. M. Sanderson, M. O. Watson, T. Leane, A. Kruys, and W. J. Russell (2009) “Monitoring with Head-Mounted Displays: Performance and Safety in a Full-Scale Simulator and Part-Task Trainer,” Anesthesia & Analgesia, vol. 109, no. 4, pp. 1135–1146.
[2] Biocca, F. and J. Rolland, (1998) “Virtual Eyes Can Rearrange Your Body: Adaptation to Visual Displacement in See-Through, Head-Mounted Displays,” Presence, vol. 7, no. 3, pp. 262–277.
[3] Krupenia, S. and P. M. Sanderson, (2006) “Does a Head-Mounted Display Worsen Inattentional Blindness?,” Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 50, no. 16, pp. 1638–1642
Google Glass showcases London <http://www.dailymail.co.uk/sciencetech/article-2665865/Google-Glass-launches-UK-Explorer-scheme-open-British-adults-headset-set-1-000.html#v-3637689178001>
Prostheticknowledge. Autodesk Maya 3D HoloLens Demo at WPC. Tumblr. <http://prostheticknowledge.tumblr.com/post/124091566526/autodesk-maya-3d-hololens-demo-at-wpc-microsoft>