Which are most promising new market sectors for service robots over the next 10 years in your view?
We see plenty of opportunity in our existing markets, particularly logistics. Today there are only around 15 service robot companies targeting the logistics market, but even if there were 100, there would still be a huge market opportunity. Remember that traditional industrial robot companies have had 50 years of market growth in manufacturing, with no real pressure to diversify to other sectors. We’re now seeing collaboration with industrial robot manufacturers in newer sectors like logistics, where they supply the robot arm, and service robot manufacturers provide mobility and workflow planning.
The other market where I see a big opportunity for service robot manufacturers is in elder care. There are so many potential opportunities for using service robots to assist overburdened healthcare workers in fetching, carrying and load bearing. But there are technical hurdles that we need to address. In particular, we need sophisticated sense-and-respond technologies when dealing with fragile patients. For example, if a patient leans on a robot for support, the robot needs to be able to sense that pressure and not start to move until it has gone.
In Asia, we see companies focusing on applications for direct robot-patient interaction. I think there is greater cultural resistance to that in the US and Europe so in these markets we’ll see more service robots being used to support caregivers rather than interacting directly with patients.
Which technological advances are you most excited about in terms of meeting customer needs?
There are two areas of development that are especially important. One has to do with enabling humans and robots to work together in increasingly close interaction and collaboration. The other is business models that enable more companies to adopt robots, and robot providers to offer additional value-added services to their customers.
Many customers want a very fast-moving robot that might be working with heavy weights, that is still collaborative. Those capabilities have historically not been compatible – as soon as you have a very fast-moving robot, or one working at speed with heavy parts, it has been caged. That’s starting to change and will continue to do so, thanks to improvements in sense-and-respond technologies. The sensor market is moving at a rapid pace – I saw at least 15 new laser companies at last year’s Consumer Electronics Show, for example. Manufacturers of robot arms are increasingly building sense-and-respond capabilities – power-force limiting, torque-sensing – into the robot arm and developers are working on tactile capabilities for grippers and other end effectors.
We’re going to see more complex applications that combine data from multiple sensors in real time – from 2 or more robots, plus external sensors in the warehouse or factory, for example – and use sophisticated algorithms to enable the robot to respond accurately to their environment. We’re gradually stretching the envelope from robots traditionally only working in very structured, predictable environments, to robots working in semi-structured environments with workers, for example in manufacturing cells. Having robots work directly with humans in very unstructured environments – for example in construction, helping a worker lay tiles in an irregular pattern – is further off. The technologies for manipulation and dexterity are there, the issue is how to get the robot to do the right thing at exactly the right time.
The second area I think is very promising, and which we’re investing heavily in at Fetch, is Robots as a Service (RaaS), where customers rent a robot rather than owning it. This frees up valuable capital. It also means the company doesn’t have to hire technicians to maintain a fleet of robots from a number of different manufacturers with different technical requirements. The RaaS model is timely because many large companies currently have ageing equipment, including robots, that they are thinking about replacing. For smaller companies, RaaS offers an easy entry into robotics, with low training and maintenance costs, and no sunk capital.
I would predict that over the next 5 to 10 years, we’ll see up to 70% of professional service robots in the market being acquired by customers in a RaaS model, particularly among small-to-medium sized companies. We’re seeing the RaaS model take hold in the industrial robot market too, but I think development there will be slower as the ownership model is more heavily ingrained with large manufacturers.
The combination of these two trends – more sophisticated sense-and-respond capabilities and a cloud-based robot service model – enables service robot manufacturers to provide services that add value beyond the return on the robot doing the job it was hired for. I see three areas for added-value services. The first is in productivity improvement – analysing robot data and data from other sensors to analyse and improve workflows. The second is in safety. For example, we can perform near-miss analysis that takes data from multiple sensors and analyses it to see where a collision was only narrowly avoided or avoided at the cost of machines having to slow down. The third area is in auditing and compliance. Service robots are already used in retail to audit stocks and ensure shelves are filled. But there’s also an opportunity for auditing customer moods. Many service robots in retail interact visually with consumers through video or photos. Imagine the opportunities for improving store layout and service if that data could be analysed to analyse a customer’s mood as they move through the store – what delights them, what annoys them. Finally, the data collected through sensors can also be used to ensure compliance – for example, that fire-extinguishers are in the right place and fire doors are not blocked.
What are the challenges in enabling these added-value services within a RaaS model?
Customers are mostly concerned about cyber-security in a cloud-based model, but the large cloud storage vendors have by now gained significant experience and use sophisticated cyber-security tools. Actually, the highest security risk is on the customer side is in managing passwords safely, for example though 2-factor authentication.
For the robot service provider, the biggest challenge is getting the data off the robot as more and more sensors are added. At Fetch, we’re investing in algorithms and compression techniques to deal with that. Currently, robot providers work with the data extracted from the robot to provide services to customers. I think we’ll see the rise of robot infrastructure companies that offer robot fleet management and data management, who will also provide data analytics toolkits to customers to do some of the work in areas like productivity improvement and workflow analysis themselves.
What are your goals as Chair of the IFR Service Robot Group?
I’m hoping to focus on providing members with reliable information on trends and markets that will enable them to make good investment and business-strategy decisions. Service robotics is a very broad field and it’s difficult to collect market data covering all manufacturers and application areas. I think the IFR has the competence and reputation to offer reliable, up-to-date market information to members which means we can engage more service robot providers in providing the data needed to make these statistics comprehensive across the service robot sector.
Picture: ? Fetch Robotics