Can AR Help Manufacturing?

I found this interesting video from Microsoft about how AR could be used in a manufacturing environment, and specifically in the car business.  Currently, I am a Manager of Quality and Engineering at a Toyota facility here in the US, so I have my own take on what AR can do for manufacturing.

The video goes through a variety of stages of car design from test driving to assembly to purchase.  While I don’t have intimate knowledge of every stage of the process, I’ve seen enough of it to know that this video idealizes the usefulness of AR.  I also realize that the point of the video is to stretch the imagination, and not to give a point-by-point improvement plan of AR, but I’d like to clarify the reality anyway. 

I am an AR enthusiast, but I don’t want people to have an unrealistic expectation of what AR can do.   One area I thought glossed over the reality was in the design phase.  The video suggested that the AR desk would magically transform the sketch into a design.  The reality of design is it takes thousands of manhours to get every detail right.  It’s extremely important to make sure all the dimensions add up, and not interfere with other parts.  Four or five years before a part is actually made in production (and I’m speaking of the more complex parts), I’ll get a print or a 3D model to understand how it will affect the production.  So I don’t think AR is going to suddenly reduce design time.  Design time improvements would come more from linking together various 3D models easier which is more about processing speeds and point-clouds. 

I also disagreed with the idea AR would help with maintaining equipment, as shown when the woman walked by the robot and scheduled maintenance.  Computerized maintenance systems are the same as any database in the world.  You get what you put in it.  We find simpler paper systems work much better than elaborate computerized ones. 

Areas I think AR would help would be in inspection and manual assembly areas.  The exact manner a person puts a part on matters greatly in the final quality of the part.  By using AR to give a visual guide to the worker, the worker is more likely to do it correctly the first time.  And if we use the idea of “games” from yesterday’s post, then we can help the worker understand how well he’s doing and keep score for the day. 

Inspection would also be easier with AR because the specifications could be seen easily.  Also, judgement of defects would become easier as AR could project measurement scales onto parts.  In casting plants, its often difficult to decide good or bad of certain types of defects because the exact size is hard to discern.  This is compounded by a short inspection time.  AR could help the worker quickly make a decision. 

The other area I think AR would help with is information sharing.  I think part design is too specific and detailed, but general schedules, production progress or other non-critical information would be useful in an AR environment. 

So my conclusion is that, yes, AR can help manufacturing, but it’s not a panacea to be applied to all problems.  Useful applications of AR could immediately improve manufacturing, but only if applied in ways that are trying to solve an existing problem and not just play with pretty toys.