What I Learned: Mechatronic Product Development and the Talking Refrigerator
What I learned this week … came from the keynote and press conference at IBM’s Rational Software Conference (RSC2009). IBM is talking about how to help companies develop and manage today’s smarter products. 
What was surprising to me is that the conference is focused on developing software - not physical products - but that a lot of the conversations focused on manufacturers and product development. Are we finally getting to the point where ALM (application lifecycle management) and PLM (product lifecycle management) can be discussed in the same sentence?
A Little about “Mechatronic” Products
According to IBM’s keynote, 70% of products have embedded control systems. This means that your next refrigerator may just be “smarter” than your first PC. OK, that really depends on how old you are, but my first PC wasn’t all that smart. The point is that a traditional mechanical product has evolved to incorporate a significant amount of software. Engineering and product development has evolved from mechanical design to a combination of mechanical and electrical design (like a “dumb” refrigerator) to mechanical, electrical, and software design in more sophisticated products (like a smart, talking refrigerator that automatically adjusts itself based on usage, season, time of day, or other factors). Another statistic quoted was that 90% of all innovation in the automotive industry is in software. While I can’t validate the percentages, the sentiment is definitely true. Many products today would not be what they are without software.
So What’s the Problem with Mechatronic Design?
There are three very distinct worlds within product development:
- Mechanical Design - The physics of the product. For the refrigerator it is the body, the handles, the shelves, the compressor (or parts of it), and other physical aspects of the product.
- Electrical Design - The electronics in the product. This can be as simple as wiring, more complex like a printed circuit board (PCB), or maybe a fully programmable chip or processor (that in turn requires software).
- Software Design -The brains of the product. This can included software algorithms that are embedded on the chips of the product, or could include programmable functions of the product. Hint - ever notice that some of your products you didn’t expect to hook up to your computer have a USB port? It might just be an indicator that your product (perhaps your tv set today, but maybe a lot of other products in the future) is set up to get software updates from the Internet.
So why is this a problem when developing products? The fact that there are three distinct design elements of a product is not the problem. The problem is that each of these design elements has it’s own lifecycle, and each impacts the other. If the mechanics, electronics, and software were unrelated then they could all be nicely designed in parallel without issues. Unfortunately, what makes today’s products “smart” is exactly what makes them hard to design and manage - the software is a key part in controlling how the product’s electrical and mechanical elements function.
Implications for Manufacturers?
The implications for manufacturers today is that product design is getting more difficult (as if it wasn’t hard enough). Processes like change and configuration management that are already hard for one discipline (mechanical, electrical, or software) need to be elevated to the systems level to encompass the whole product. Teams working on individual aspects need to collaborate earlier in the design process.
This will not happen overnight, but the companies that get this right will have a tremendous advantage in bringing high quality products to market, and avoiding late conflicts between the different disciplines that drive high product development cost and product introduction delays. This is the future of product development, and today’s disjointed processes will not be competitive when the leading companies figure this one out.
So that is what I learned this week, I hope you found it interesting. I will post later this week on what I hear from IBM in regards to addressing mechatronic design issues, and what their vision is for addressing ALM and PLM holistically. Let me know what you think.
NOTE: OK, this picture is not a real talking refrigerator, I admit it. This is a toy. But toys are just one more example of mechatronic products, and they will continue to get more sophisticated (incorporating physical motion, Internet connectivity, and “thinking” over time).
Marc commented:
I think your analysis is right on target. The three discpline (electrical, mechanical and software) coming together will be a killer application for the future of intelligent product design. Nowdays many product are being designed in virtual/simulated environment. Electronic has done it for year, Mechanical for even longer. Unfortunately all have been designed in silo. But I think things are changing as it comes to electronic and software. Software runs on processor and new technology enables now virtual modeling of the hardware at simulation performance that perfectly meet the need to the software developer (i.e in the pas electronic simulation was done at a detail of the hardware that was not relevant for software). Some companies are putting this technology as Electronic System Virtualization (create a virtual model of your hardware on which you can run the actual embedded software) One company example is CoWare www.coware.com . This type of simulation should then be connected to mechanical type simulation to offer design a real completely simulated environment that can keep these teams synchronized.
Jim Brown, Tech-Clarity commented:
Vinod,
I love your question! I think the answer is "no." The lifecycle of your washing machine will likely be based on the physical parts wearing out. Unlike your computer, your washing machine isn't likely using every bit of computing power available. I don't see turning in an old washing machine because it won't run a new application. What I am saying is that your PC's performance is dictated by computing power. Your washing machine probably is not.
Now, there are two other important perspectives:
1 - There is the opportunity to extend the lifecycle of your washing machine with smart technology. We are not there yet (heading there in cars, planes, industrial equipment, and even industrial lighting from a recent conversation), but the ability for equipment to self-monitor and report a service problem might extend the life of the unit. If your washing machine can e-mail or text message you (Twitter maybe?) and let you know it needs a new bearing, maybe it gets fixed before you burn out the motor. That could extend the lifecycle. Wishful thinking? Maybe today. But the "internet of things" concept - that we will eventually connect everything together so it can all communicate - might prove to extend the lifecycle.
2 - Having said the lifecycle might be extended, it might also be more difficult. You will have a more complex washing machine. If there is a bug in the code, your machine might break. And it might be difficult to diagnose and fix. You might need to download new patches for your washing machine so you can keep your whites white and your colors bright! On one hand it offers the ability to fine-tune performance over time, on the other it offers the potential for more things to go wrong. So you might have to reboot your washing machine. You might have to install new software. Today, it is very simple code and very well tested. But it will get more complex. Let's just hope it gets networked so it can download its own patches. And lets hope nobody can download a virus onto your washing machine to shrink your favorite sweater!
So it will be more complex, might last longer, but also has more things that can go wrong with it. I am sure there will be debate over how automated planes will be after the recent Airbus tragedy. How much should we allow computers to be in control? Definitely trade-offs.
And - it makes managing product development harder! But smart products are the future, I have no doubt.
Jim Brown, Tech-Clarity commented:
Oleg,
Yes, ALM and PLM are very different - and very similar. I don't think we will see a process (or tool) that works equally well for both. To me, the model that will work is to have coordinated processes that manage the work in each area. I believe this is also true for tools. I can't envision one "ubertool" to use for developing software, developing electrical designs, and developing physical designs. I do think there needs to be a systems model that pulls those three together, and there is a lot of commonality in how that information needs to be used in the enterprise and managed as IP. To me, this is a new layer that sits on top of the existing processes and tools. I know that is not the direction everybody is taking, and a single elegant tool for all would be great. But to me, I think you need to enable each set of innovators independently and then add a layer of coordination on top. Thoughts?
Vinod Kumar commented:
An increasing use of software or "intelligence" embedded in products is evident and future products are bound to more intelligent. But the fact remains that more and more innovation in todays products is technology driven.
Now i want to bring in analogies between a home computer and a washing machine, if future washing machines are smarter than my first PC will the life cycle of washing machine be similar to my first PC?
Oleg Shilovitsky commented:
Conceptually you can say ALM and PLM is about the same. However, I think, there are some lifecycle differences in hour you manage software product vs. how you manage hardware /physical product. Software build is different from manufacturing Bill of Material in my view. But both need to be combined together.
