Major Takeaways:

  • A Boeing internal study showed that augmented reality significantly boosted trainee performance when asked to assemble a mock wing for the aircraft manufacturer
  • Trainees armed with AR-enabled instructions were 30% faster and 90% more accurate on their first tries than other groups who did not have access to augmented reality instructions
  • Workers more easily trusted instructions delivered through augmented reality compared to static PDF-based instructions

From the Re/code report:

A big hurdle for convincing companies to try augmented reality, speakers said Monday at the Augmented World Expo in Santa Clara, Calif., is the lack of data and historical models proving that AR works.

However, universities and a handful of big companies are starting to gather that data. In a presentation, Boeing electrical engineer Paul Davies shared the results of an internal study of how trainees performed when asked to assemble a mock wing for the aircraft manufacturer; the study’s subjects were sorted into three groups and could refer only to the instructions on a PDF displayed on a desktop computer, a PDF displayed on a tablet or animated AR instructions displayed on a tablet.

The AR-tablet group were 30 percent faster and 90 percent more accurate on their first tries than the other groups, Davies said. If AR continues to prove similarly effective in continued research, it could be “huge” for companies with Boeing’s scale, he added, both in terms of increased quality and decreased cost.

One factor contributing to that speed: The augmented reality trainees were more willing to quickly accept what the instructions were telling them to do. One step called for a group of washers to be placed together without spacers, which tripped up the groups consulting a static PDF.

A big and already well-known challenge for AR, however, is “occlusion” — when and how virtual objects block out the real world. Davies said the most difficult step for the AR group was one that superimposed a virtual piece of the wing over the tablet’s view of the real piece, when it would have been easier to understand if that virtual object had appeared behind the real one.