Six ways robots are mainstream and four ways they are not

At a university like MIT there is advanced technology everywhere. I cannot walk or talk without stumbling into new, mind blowing gadgets, sensors, software, and, yes, robots. The question is what does this mean for robotics outside MIT. Is this finally the time for robots to go mainstream or am I being duped into thinking so?

The first industrial robots began their life in 1938 with a programmable mechanic crane, but the first robotic patents were filed in 1956 and robots have only been in true industrial use since the mid-1970s. Like earlier technologies, robotics has arguably evolved from clunky, experimental, and dangerous to elegant, industrial, and harmless. But what is the evidence that robotics now is beyond high tech extravagance and is moving into the realm of the industrially and practically sound?

Ornitopher from the Tedrake lab at MIT. Photo credit: Jason Dorfman, MIT.

  Digital robotics is a mature wave of automation

According to my colleagues Brynjolfsson & McAfee, after the automobile, railroad and aircraft went mainstream last time around, recent digital advances now puts us in the Second Machine Age (2014). At MIT, this takes the form of hybrid innovation—hardware guided by software interfaces and with sensors embedded. For sure, there is plethora of robotics program around campus, doing everything from the extravagant to the mundane. Cynthia Breazeal’s emotional robotics at the Media Lab carries the promise of robots contributing to happiness. Julie Shaw’s interactive robotics improves ways that robots can work alongside humans. Russ Tedrake’s robot locomotion group can make a robot land on a perch just like a bird. There are in fact 30 such programs at MIT in total, according to the Robotics@MIT website (all are listed here).  

Robotics research is maturing--companies are spun-out and acquired

The US National Robotics Roadmap, first launched in 2009, brought together all the national technical university hothouses under one strategic umbrella and got the government involved through DARPA robotics challenges slightly reminiscent of the early days of the Internet. The 2013 roadmap focuses on the manufacturing, healthcare, service, and space industries.  With Google and Facebook starting to acquire robotics companies that used to be small university spinoffs, bots are back in the media, and I’d be willing to bet most people could name as many fictional robot characters as they could name Hollywood actors. In Massachusetts, the MassTLC Robotics Cluster, run by The Mass Technology Leadership Council (MassTLC) technology association raises awareness for and spurs growth of the region’s thriving robotics industry, and MassRobotics, a new robotics hub with a breeding ground of some 100+ companies, is being launched, spearheaded by Daniel Theobald, CTO of the robotics and healthcare company Vecna. According to MassVentures (2014), over the past two years, 29 percent of US robotics deals happened in Massachusetts with Silicon Valley as the runner up at 18 percent.

Robots are more versatile, cheaper, and easier to operate 

The multitude of use cases for the new generation of robots—across industry—alongside humans—and as a service army for repetitive, strenuous, or labor intensive tasks in our daily lives as well as in dangerous situations—in emergencies, on the battle field, and as proxies for human contact—makes them more and more attractive. However, even quite advanced features, are now moving into the prosumer range. The example is a quadrotor—a small helicopter drone that can carry out useful transport and surveillance tasks, is easy to build and costs less than $5000. That being said, the high end, humanoid Atlas robot from Boston Dynamics, poised to become the next soldier or even Ebola first responder, still has a Ferrari price tag. However, as each of these markets become more understood, mass production will kick in, and the prices will come down even further. 

  The Europeans are deeply into robotics--and the Asian market is buying 

Robotics is obviously not a purely US preoccupation. According to the Brussels based non-profit euRobotics, from  today’s  €22bn  worldwide   revenues,   robotics  industries   are   set   to   achieve   annual   sales   of   between   €50bn  and  €62bn  by  2020, and  Europe has   a 32%   share   of   the   industrial   market. 24-30 November 2014 is European Robotics week, since “Robots are able to capture our imagination like no other tool”. Swiss multinational ABB is a pioneer and giant in industrial robotics with a 190,000 robots install base and 1500 professionals in 45 countries dedicated to robotic applications. For comparison, you will recall that Ford sold 15 million black Model T cars from 1908 until1926 (although 4-5 colors were available from 1908-14, they were not popular). ABB recently and symbolically changed its flagship IRB 1600 robot from the warning color orange to the more dynamic color graphite white, although you can apparently now order robots in any color you want. According to World Robotics survey on industrial robots (2014), more than 200,000 industrial robots will be installed in 2014 worldwide, and Asia was by far the biggest region with 98,807 industrial robots sold, 17% higher than in 2012, hitting the highest level ever recorded. By 2017, they project more than 400,000 industrial robots will be installed in the factories of China.

Photo credit: Jason Dorfman.

  With more advanced sensing, robots can now work alongside humans 

For decades, robots have been part of the avant-garde of technology, to the point that they actually were considered almost mundane for a while, at least by the general public. In industry, robots were efficient in a factory setting, but also somewhat dangerous, and confined to set tasks behind closed gates, far away from humans. In 2014, the picture is different. Since 2002, iRobot has sold more than 8 million home robots, most famously, Roomba, the autonomous robotic vacuum cleaner. In the lab, new, agile robots are deployed “everywhere”—they are safer to be around—and with that feature alone moving out of the comfort of the lab, they are set to transform industry but also society.  

The new generation robots, which benefit from advances in sensor and motion technology, and vastly more complex computational algorithms, are designed to interact with humans, move freely around, and the use cases for such new types of robots are vastly different than was the case even just a decade ago. The challenge is, of course, that to get here, an enormous investment has been made by the robotics pioneers. So, the question is, will it take the same amount of investment to bring all this excitement to the mass market, and if so, who will do it? Typically, any mainstream market is fueled by consumer demand, not industry applications. Even with advanced sensing on board, will robotics do one, both or none?

Robotics is poised to become a platform of innovation

With all the technological progress, the lower price points, and the accelerating use cases, one would think the market is open for huge growth. True, but the question is at what magnitude and speed. In order for robotics to truly become a platform for innovation in the order of the Internet (for lack of better metaphors), and in order to achieve sufficient scale to do so, new physical, shared infrastructures and tools are needed, and involving a new set of talent from the business development side of the equation.

Four platform innovation requirements still remain before robotics go mainstream; critical mass, interoperability, deep industry engagement, and more robust testing infrastructure.

Platform innovation, while appealing, is difficult to pull off. Here are some factors that are needed to bring robotics into the mainstream.  

Critical mass

Arguably, ABB, the Swiss multinational, and KUKA, the German €1.7 billion (2013) industrial robotics maker, Boston Dynamics, now owned by Google, Kiva systems, now owned by Amazon, and iRobot, still going on its own and which just hit the $0.5Bn revenue mark, all fit the bill as robotics success stories, but this is not enough. Up and coming Harvest Automation with its HV-100 agricultural robots and Vecna technologies with its QC series healthcare robots will move the mark. But individual robot and robotics company brand recognition needs to move by leaps and bounds before robots could be considered mainstream.  To the extent Google or Facebook truly become known for robotics, that would change the game, but those brands are so far known for so many other things as well, that robotics doubtfully would be their signal attribute. To a certain extent, the killer application for robotics has been missing, and it is not certain that we should still be looking for one, since the impact might be more subtle and lasting if robotics is allowed to more seamlessly integrate into manufacturing, without being expected to always have a wow factor.

Another measure of going mainstream is the ever trend hungry venture capital, and there, robotics is still small potatoes. According to Hizook (2013), annual robotics VC is growing but is still only at the $250M mark. That being said, the National Venture Capital Association (2014) does report that robotics showed the biggest year-over-year increase in global investor confidence.

For robotics to take off, using robotics technology must become an acceptable part of business culture both in terms of cost/benefit, risk, and habit. Deep integration of robots in to the manufacturing process brings operational challenges that demand a new type of operational research and a highly skilled workforce that not only can operate advanced robots, but who can model and predict advanced manufacturing patterns. We also need a critical mass of providers, original equipment manufacturers (OEMs), as well as customers. Is the current cohort of robotics startups coupled with a few large industry players (such as ABB), enough to provide that backbone? No industry giant has yet become known for actively promoting a robotic production environment.

A shared, interoperable technology platform 

Without the ability to interconnect, the various robotics technologies and indeed each of the robots out there could neither really start interacting amongst themselves, which they need to do not to run into each other, nor could they interact with humans. Currently, the top robotics software out there is proprietary, and open standards are in their infancy. Some open source software exists, most prominent is the EU built iCub.org platform, and from the DARPA challenges, but a lot of hard work remains to convince the industry (and researchers) to embrace it, or perhaps simply to further robot interoperability standards. The evolution of ISO 184/SC2’s standardization work on Robots and robotic devices, including ISO 13482 on personal care robots is great. However, standardization of even one feature typically takes from 2-10 years depending on the complexity of the standard and the amount of convincing it takes. In a rapidly evolving environment this poses challenges.

Robustness and cross-industry engagement  

We now have plenty of industrial use cases out there, from factory floors, to intelligent cars, to domestic uses around the house. Indeed, industrial robots are starting to perform very well in complex tasks. However, there still is a gap between what can be shown in the lab and what can be put to use and that gap needs to be closed before the industrial attention span again closes robotics off to simple automation or simply to a science fiction state. As Prof. Russ Tedrake at MIT points out, industrial robustness is different than being able to showcase something one-off on YouTube. You need formal verification, compliance to safety and other standards (many of which still yet to be formulated), and shared measurement methodologies. In order for that to happen, there must be a partnership between government, the advanced manufacturing industry, R&D institutions, and startups. They need to develop common standards, rapid field testing, and build awareness throughout industry that robots should be given a tryout chance in completely new settings. This is not done through sensational focus on gadgets, but rather on focusing on 24/7 reliability and perhaps, from the point of view of the media or the enthusiastic roboticist, on somewhat boring, industrial use cases.

Physical testing infrastructure

The US robotics community is large and growing and consists of a myriad of polytechnic universities (such as MIT and Caltech) and a cluster of other research institutions and corporate R&D labs and startups. What is sorely lacking is a testing ground where all of the possible robotics applications can be tested in full scale, whether this is true to life city settings or industrial scenarios. The community of Devens, MA, is proposing itself as a somewhat unique “interoperability playground” where robots can be tested, but more sites are needed, and these would need to be publicly funded and shared to have a platform effect. 

Is robotics going mainstream? We have established that robotics is the name of the game at MIT, which is step one. The Boston area also has a critical mass of smaller startups, which helps. The state of Massachusetts, in fact, is a hothouse for the kind of advanced manufacturing that robotics requires, although there is no statewide robotics strategy.  The US robotics roadmap and the DARPA challenges, the euRobotics effort, and the increasing sophistication and purchasing power of Asian manufacturing all contribute.

In the end, mainstreaming robotics will take more than technology. It will take more than organizational effort. What it takes is proven performance over time. 

Tomorrow’s robot next door will likely be more mundane, akin to Roomba, the vacuum cleaner, than approaching singularity with legs and locomotion like Atlas and will not have a brain like HAL 9000, the fictional character in Arthur C. Clarke's Space Odyssey series. But if you can work happily alongside robots, you are embracing the future. I will be telling that to my kids, 8, 6 and 2, and so should you.

Trond Undheim is a Senior Lecturer in Global Economics and Management at the MIT Sloan School of Management, leads MIT ILP’s Startup Exchange (STEX), connecting industry to startups, and is the Founder of Yegii.com, the insight network, a strategic consulting firm for the digital age. Follow him @trondau