This is a survey of potential industrial use cases that spans a variety of domains. The article feels like a submarine [0].
It's hard for me to imagine getting from where most businesses are now (unable to share data internally among teams, limited investment in web dashboards that show live data) to a fancier "live data overlaid with physical systems." Is there that much value differentiation? I get the tech appeal, and it looks cool. I'm not clear the cost/benefit is there in general.
I do suspect AR will be adopted in two scenarios. In niche domains where cost/benefit is acceptable (where?). And the other scenario is "shiny buttons and blinky lights" pet projects.
I'm the writer, and I can relate with your opinion on whether the cost/benefit is there. From the responses, it appears many wonder this as well.
The biggest benefit I can imagine where the cost is outweighed, is when AR is used to reduce downtime of huge output facilities. If using AR could save an organization even just 10 minutes of downtime in a year, depending on their output, it could likely be worth the investment.
Time will tell how it greatly it gets adopted. We sure as hell are betting on the fact that it does. And now seems like a good time in the adoption curve to place that bet.
I think AR is being used lightly in robot-assisted surgery and in the future telesurgery. But to my knowledge so far AR is used more for the trainings on how to use the machines. The surgeons already work with sort of massive stationary VR glasses, so it is not that big of a jump to augment the image it if it adds value.
Yes. For it to be useful, you need a detailed, accurate model of the thing you're working on. They exist for aircraft because they're needed to build the thing. Before 3D CAD, full-scale mock-ups were built to figure out how everything fit, and 3D CAD is cheaper than building a full-scale wooden mock-up.
Few other objects are that big, that complex, and have a sufficiently detailed model available. For a building, updating the model is a lot of work and requires discipline to keep the model in sync. Nuclear plants, spacecraft, and unusually complex chemical plants might have a model at that level of detail.
Interesting. Would love to chat more and see what you've built and show you what we've built - shoot me an email if you get some time - sean@aircada.com
I suspect nobody. There was this whole push to create "enterprise" use cases for VR once the current crop of startups realized the consumer market would never take off. A lot of them were successful at selling their useless prototype "products" to company executives, who I'm sure made unilateral decisions to implement the devices with no ask from engineering. But the technology is light years away from being practical for daily use. See: the entire MS/DoD HoloLens debacle.
I see your point as we own a Hololens2 and quickly realized that yes, the wearable tech is still light years away from being practical. What's not light years away, however, is the device that everyone has sitting in their pocket.
If you can pull out your phone, point it at any asset, and immediately see equipment specs, SOPs, and guided procedures, there certainly is value in that. If you're trying to find pump a32 in a massive facility, and can type it in your phone and be directly guided there, surely there is value in that as well. These use cases go on and on.
Coarse-grained systems like that are probably more valuable in the near term.
Especially if you can take a picture of pump a32 and its asset tag, and that picture and location goes into a searchable database without much typing. That lets you build the database over time.
If everyone used Hololens2 for this now and then the smart phone came about tomorrow it would be seen as a huge upgrade with not having to wear the headset.
Building services engineer here. I think AR could be useful for "x-ray" views of services above ceilings and in risers where you're trying to locate concealed equipment. When it's sitting in front of you in a plantroom you really just need to look at it or have a regular screen to tell you what's going on in terms of operations.
The big assumption is that the 3D model of equipment used by an AR app matches what's actually installed on site. This is rare in my experience - particularly over time as maintenance is performed in bits and pieces without documentation being updated. It's exceedingly rare to find an old building with accurate documentation of where everything is. I'd love an app that could compare a laser scan of installed plant with the 3D design model and automatically align them into an "as-built".
It’s also impossible in the vast, vast majority of real-life situations.
As DeNiro said in the terrible
Movie “Ronin” — “But the map…the map is not the territory.”
The 3D model will differ from the actual construction. From the very beginning and more and more over time. Documentation is hard enough for software projects. Why would we expect documentation quality of construction to be better?
I can see this being used for retrofit designing. Use a LiDAR point cloud to create a model of an existing open facility, all the exposed machinery and piping. Then suggest changes for future installation and have current staff walk through the plant checking the virtual future piping/etc.
We already do that without the 3D glasses though. Just 3D rendered to computer screen.
> The 3D model will differ from the actual construction. From the very beginning and more and more over time.
Not really. The point of construction is to turn the "3D model" into reality.
Some problems in the original planning can and will be spotted during construction but that just means plans are updated once conflicts are resolved.
There's even a occupation specialized in ensuring that what's being constructed complies with the "3D model": construction and building inspectors.
The mental models that are formed by software developers based on their personal experience within software development activities is not compatible with any day-to-day practices of any engineering field.
> Documentation is hard enough for software projects. Why would we expect documentation quality of construction to be better?
Because it is, by it's very nature and purpose. Practically all of the engineering deliverables are written documents. These documents held engineers liable to any defect. In engineering, documents matter the most.
Engineering is not software development, and software development is not engineering. Engineers need to put in writing stuff that works for decades, at their financial and even civil and criminal risk. You don't just hit the "build" button and retry if something failed. You put into writing exactly how something needs to be done in order to meet all requirements, and make sure that everything is done according to what was put in writing. If anything goes wrong, what's written down is the engineer's only defense.
> There's even a occupation specialized in ensuring that what's being constructed complies with the "3D model": construction and building inspectors.
Respectfully, I’ve done this specific job in the past for industrial/commercial building construction. It’s very, very difficult to get builders to build the thing to perfectly match the drawings. They often build what is convenient, or make absent-minded changes.
We do try to update the drawings to match reality (a revision known as “as-builts”) but perfection is very elusive IRL.
I was a chemical/electrical engineer by training and trade before moving into software engineering.
>> Documentation is hard enough for software projects. Why would we expect documentation quality of construction to be better?
> Because it is, by it's very nature and purpose. Practically all of the engineering deliverables are written documents. These documents held engineers liable to any defect. In engineering, documents matter the most.
Also, while you can make changes, you can’t really refactor a building or migrate it to a different service.
I think actually the problem would be placing the model accurately in a general indoor case.
No GPS, not that that is accurate enough anyway. There are other tricks to do it, restoring from a known point say.
When used in industrial environment, AR will use indoor positioning systems to make sure the model and real world are aligned.
Would love to chat more with you about this. Our app could definitely provide value in construction and we'd be very open to a pilot program on your terms. In all honestly, we are new, so feedback and an ongoing relationship are the biggest things we're looking for at the moment.
I can't see a device that displays graphics between a person's eyes and any feature of an industrial plant being very popular with said industrial plant's lawyers.
I think there is a misconception that the 'metaverse' can be some kind of data driven visual experience that we immerse ourselves in.
The real metaverse is what lives inside our heads - that we communicate with each other to (hopefully) build a shared mental model with enough concordance with real reality to be useful to us. All media (including AR) is no more, and no less, than this communication.
31 comments
[ 4.2 ms ] story [ 72.7 ms ] threadIt's hard for me to imagine getting from where most businesses are now (unable to share data internally among teams, limited investment in web dashboards that show live data) to a fancier "live data overlaid with physical systems." Is there that much value differentiation? I get the tech appeal, and it looks cool. I'm not clear the cost/benefit is there in general.
I do suspect AR will be adopted in two scenarios. In niche domains where cost/benefit is acceptable (where?). And the other scenario is "shiny buttons and blinky lights" pet projects.
[0] http://www.paulgraham.com/submarine.html
The biggest benefit I can imagine where the cost is outweighed, is when AR is used to reduce downtime of huge output facilities. If using AR could save an organization even just 10 minutes of downtime in a year, depending on their output, it could likely be worth the investment.
Time will tell how it greatly it gets adopted. We sure as hell are betting on the fact that it does. And now seems like a good time in the adoption curve to place that bet.
https://upload.wikimedia.org/wikipedia/commons/3/34/Robot_as...
Few other objects are that big, that complex, and have a sufficiently detailed model available. For a building, updating the model is a lot of work and requires discipline to keep the model in sync. Nuclear plants, spacecraft, and unusually complex chemical plants might have a model at that level of detail.
If you can pull out your phone, point it at any asset, and immediately see equipment specs, SOPs, and guided procedures, there certainly is value in that. If you're trying to find pump a32 in a massive facility, and can type it in your phone and be directly guided there, surely there is value in that as well. These use cases go on and on.
Willing to chat about this more? =]
Would love to pick your brain a bit!
Thanks!
The big assumption is that the 3D model of equipment used by an AR app matches what's actually installed on site. This is rare in my experience - particularly over time as maintenance is performed in bits and pieces without documentation being updated. It's exceedingly rare to find an old building with accurate documentation of where everything is. I'd love an app that could compare a laser scan of installed plant with the 3D design model and automatically align them into an "as-built".
As DeNiro said in the terrible Movie “Ronin” — “But the map…the map is not the territory.”
The 3D model will differ from the actual construction. From the very beginning and more and more over time. Documentation is hard enough for software projects. Why would we expect documentation quality of construction to be better?
I can see this being used for retrofit designing. Use a LiDAR point cloud to create a model of an existing open facility, all the exposed machinery and piping. Then suggest changes for future installation and have current staff walk through the plant checking the virtual future piping/etc.
We already do that without the 3D glasses though. Just 3D rendered to computer screen.
Not really. The point of construction is to turn the "3D model" into reality.
Some problems in the original planning can and will be spotted during construction but that just means plans are updated once conflicts are resolved.
There's even a occupation specialized in ensuring that what's being constructed complies with the "3D model": construction and building inspectors.
The mental models that are formed by software developers based on their personal experience within software development activities is not compatible with any day-to-day practices of any engineering field.
> Documentation is hard enough for software projects. Why would we expect documentation quality of construction to be better?
Because it is, by it's very nature and purpose. Practically all of the engineering deliverables are written documents. These documents held engineers liable to any defect. In engineering, documents matter the most.
Engineering is not software development, and software development is not engineering. Engineers need to put in writing stuff that works for decades, at their financial and even civil and criminal risk. You don't just hit the "build" button and retry if something failed. You put into writing exactly how something needs to be done in order to meet all requirements, and make sure that everything is done according to what was put in writing. If anything goes wrong, what's written down is the engineer's only defense.
Respectfully, I’ve done this specific job in the past for industrial/commercial building construction. It’s very, very difficult to get builders to build the thing to perfectly match the drawings. They often build what is convenient, or make absent-minded changes.
We do try to update the drawings to match reality (a revision known as “as-builts”) but perfection is very elusive IRL.
I was a chemical/electrical engineer by training and trade before moving into software engineering.
> Because it is, by it's very nature and purpose. Practically all of the engineering deliverables are written documents. These documents held engineers liable to any defect. In engineering, documents matter the most.
Also, while you can make changes, you can’t really refactor a building or migrate it to a different service.
No GPS, not that that is accurate enough anyway. There are other tricks to do it, restoring from a known point say. When used in industrial environment, AR will use indoor positioning systems to make sure the model and real world are aligned.
The real metaverse is what lives inside our heads - that we communicate with each other to (hopefully) build a shared mental model with enough concordance with real reality to be useful to us. All media (including AR) is no more, and no less, than this communication.