Wednesday 20 March 2013

In-Yer-Face (Interface) My Best Friend's a Robot

I have spent the last 3 days at the Innorobo event in Lyon on a mission to change the world, get rich and meet committed people. This is the place where the great and good of the robotics world, primarily from Europe (even if the Koreans won the cutest hair prize), come to show off, find inspiration and maybe a future.

There is a second event being held in Lyon in parallel. Called eu Robotics Forum, it's maybe a bit more in depth. The French Government is a sponsor, so I guess they are worthy. The organisers wanted 350 Euros to have me along, and that was enough to pay for my hotel. I am quite new to robotics - mad keen but learning my route - called SLAM for those who know. Maybe I will go next year?

I am on a start-up friendly, strict budget and staying at the Campanile close to the TGV and a short tram ride away from the conference centre. I mention the hotel only as it has an example of one of the main issues that robotics seems to face (no pun intended) today. It's about the in-yer-face interface - aka getting the best out of a product with the greatest utility and potential.

I woke early yesterday and went to shave, trying to fill the sink with an ecologically aware depth of hot water. Failure. There was an interface problem between the size of the plug and the size of the plug hole. The plug was too big (or the hole too small?).


I managed to shave in the shower with only minor cuts.

Robotics is fast becoming an area of special importance. The French government announced a package of initiatives this week to drive investment and innovation. France also boasts to be the home of Aldebaran and their famous Shanghai dancing Nao robot: but even this company is small with under 300 employees worldwide. Nice to see their CEO today out of that dark suit and playing ball with his robot (expect cute pictures soon on the site?).

The market is growing fast - supposedly a US$ 40Bn market in service robotics by 2015 (dixit GIA via eu Robotics Co-ordination Action and FP7 money). That gives room for a dozen or more players at real scale and a couple of giants! Professional service robotics is growing at some 85% a year. That's gotta be attractive! Most of these types of robots end up with the military (a post of its own, given the debate on human-out-the-loop robotics nicely captured here) or in milking parlours.

Well, we will always drink milk and, sadly, probably keep fighting. So, it's a dairy farmer or a soldier. The availability of good, empirical data is critical to making good career decisions!

The personal robotics scene looks less attractive compared to that of its professional cousins - even if every exhibitor seems to have developed a use case that fits. Sort of Corporate Social Responsibility baked into every start-up. Clearly, the current economic straits are an issue with less disposable income afforded to edutainment robots and the like. The growth here is much more modest and the associated market sizes hardly move the dial.

And as for robots that make life easier, outside of having your lawn manicured or your 4 year-old prepared for their GMAT, don't go there.

So where's the issue? What's getting in the way? So far I can see 3 big hurdles. None seems impossible to clear. The first is a negative public perception of robots; the second a set of issues around safety and product liability and the third seems to be the lack of a workable industry structure and business model.

As yet, I claim no grand insight but, through the eyes of children, emperors sometimes lose their clothes.

Robots scare us. We don't really want them in some key parts of our lives. The European Union did a large study on the public's attitudes to robotics across the 27 Member States, surveying those over 15. This work is current, published in the last 6 months.  Overall there is a generally positive view on robots - with a more than 50% approval rating and the Nordics just love them with a whopping 88% positive view - cuddling a robot on those long winter nights perhaps?

But close to two thirds of Europeans see no role for robots in mainstream education, care of the elderly or healthcare. They have a view robots can be dangerous and unpredictable, best restricted to roles where humans might see themselves potentially endangered (rescue etc) or where dull repetition is needed. This may need to change over the very near term. We are going to have over a billion people aged over 60 around in the coming years. We will need new ways to care for them, as they age further.

Are robots to be seen as slaves? For now maybe but perhaps we will need emancipation sooner and not later.

And let's also hope those slavery folks don't get too familiar with (Technology) Singularity Theory from Vinge, Van Neumann and with the more recent musings of Kurzweil and Storrs-Hall. Basically, we meet the machine cleverer than we are around 2045. If that's true, someone needs to stop Fiona getting out of her cyber-bed or at least brush up on their Spanish!

Adoption needs Acceptance and that needs Education. As one expert may have once remarked, you may be afraid of robots being allowed to help you, but think of that every time you ride in a lift!

Now what about product liability and perceived danger? Progress takes bravery. Around 150 years ago, the UK passed the Locomotive Act, restricting horseless carriages (cars) to a speed of 4mph and demanding they be piloted by a crew of 3, with one of them carrying a red flag in front of the vehicle. At Innorobo, I have seen several robots still following this legislation!

And there was a bill in Pensylvannia, thankfully vetoed in 1896, that required any car encountering a person or livestock to be stopped, dismantled and hidden behind a bush. No bushes at Innorobo!

Clearly, personal assistance robots need stringent safety regulation. This is so important. It cannot be left to individual companies and cannot be afforded by them either. Is there no European-level idea on assisted testing? New industries need support, and surely this is one area where we could make the investment at the European level. There is aligned work in ISO standardisation on this topic. I came across this upbeat assessment from Dr Chris Harper but still need to understand where we now are. Such work is worthy to be supported and accelerated. It seems too important to fail and the effort to get it right is clearly worthwhile.

Talking to exhibitors, it is the business model that perplexes me most. I attended the panel on Investments in Robotics. I heard a lot of truisms and worried at the sound of people scribbling and feverishly noting down these apparent pearls of wisdom and religious truths. I felt we'd been here before.

I heard Dmitry Grishin of Grishin Robotics explain three ways to make money in robotics - sell them, rent them or give them away and make money on the services. He made a lot of valid points - especially about affordability and beauty - but his philosophy seemed reminiscent of the mobile phone industry of the Noughties.

Robotics looks too like an industry that wants to maintain a vertically-integrated business model. My strong sense is that to grow that this industry has to find a way to split the robot body from the rest of the application/use case and to do that quickly. Clearly, open source software helps massively, but I had a bit the feeling that open source support was a rite of passage to the robotics community and not yet a real lever for success. This could just be about maturity and the academic focus of many concerns here. Perhaps, we need to be patient. They will come and build it!

I read a great book lately that sought to explain the dramatic shifts brought about by what could be called the democratisation of technology - albeit it's rather a first world view. The book is called L'Age de la Multitude by Nicolas Colin and Henri Verdier. It is strongly written, drawing on top class and recent research to make its points.

Bottom line, the authors argue persuasively that technology is cheap; this fuels an innovation and change ACCELERATION; and most importantly this innovation is external - no entity can hope to manage all the ingredients of success anymore. Innovation to be successful is becoming Open Innovation. This is where the literal and actual multitude takes centre stage.

The world is learning to code. It's seemingly the new Latin and the major Internet players are right behind it. And those building the internet are supporting too. We have a capability revolution coming fast.

To make the most of these new capabilities, robotics companies need to build clean technical architectures that focus on abstraction and virtualisation and then couple that with resilient tool kits that promote the mass adoption of robots. This is how almost all successful software companies work. A long time ago in technology speak, the rules that split platforms, applications and integration were written. Through the many changes we experience every day, these rules have stayed in place.

Do not try or expect to do everything anymore. The product architecture should support the desired business model. Keeping control of all the levers is as impossible, as it is eventually undesirable.

In this way, we may overcome many of the issues associated with fear of the robotic unknown. We will garner the capability of professionals to develop more and more relevant vertical applications and create a virtuous flywheel effect. It's amazing but it seems over 80% of Wikipedia's content is generated free and anonymously. It's mostly, but not all, about the money.

There are angels out there. Finding them? Well that's going to need a lot more than just an ordinary App Store business model.

I did succumb and bought a small personal robot. There are many other robots available but this matches my desk. Delivery is only at the end of April, so I'll need to wait a few weeks to begin my new robotic future!

I took a decision not to paste pictures of all those cute creations. We need to make robotics a serious topic.

No comments:

Post a Comment