The annual RoboBusiness 2019 conference drew a wide range of robotic technology developers, engineers, investors, and eager startups to Santa Clara, Calif., last week for three days of keynotes, panels, sessions, and demonstrations led by industry leaders and innovators.
This year’s program featured a lineup of luminaries that included Howie Choset, a professor at Carnegie Mellon University’s Robotics Institute; Adam Cantor, Principal Robotics Engineer at RSE; Doug Olson, President of Harmonic Drive; Erin Bishop, Director of Business Development at Sense Photonics; James Kuffner, CEO of Toyota Research Institute – Advanced Development; John McElligott, CEO of York Exponential; Kristen Moore, CMO of inVia Robotics; and Robin Murphy, Raytheon Professor of Computer Science and Engineering at Texas A&M, among others.
The program was truly packed this year, but here are a few of the top takeaways from the event:
#1) Robots are creating jobs, not eliminating them
Virtually every presenter made some reference to this idea during the conference, often responding to what many characterized as alarmist predictions among industry watchers and political pundits.
“There really is no consensus on what automation robotics is going to do in terms of creating jobs or not creating jobs,” said Howie Choset during his opening keynote. “But it’s worth noting that the number of jobs created in manufacturing positively correlates with the number of industrial robots that have been purchased.”
There’s a lot of loose talk around this topic, said Byron Clayton, CEO of Advanced Robotics for Manufacturing, but at the end of the day, all innovation is a kind of creative destruction. “Automation is going to destroy jobs, and it’s going to create jobs,” he said. “You need to understand how it impacts your particular industry and how it’s going to affect your employees and address it with training and establishing new career pathways. You need to assess workforce readiness versus resistance, and you don’t want [your automation plan] to create an adversarial relationship.”
#2) Innovators need to take risks to gain market traction
“How many manufacturers do you know who would see something developed at, let’s say a university, and say, yeah, I’m going to put that right on my factory floor?” Clayton asked attendees during one of his session. “You just don’t see that. They’re waiting for that technology to get to a point where they feel pretty warm and fuzzy about it before they implement it, because if they make the wrong decision and put in technology that doesn’t work, they can not only lose money, but they can damage their existing production. It’s a big financial and operational risk that most companies are not willing to take. So, a lot of sectors stall when it comes to new technology, particularly robots. The main part of the market doesn’t take it up until way down the road.”
One way to get in sync with the risk concerns of potential buyers, Clayton recommended: don’t fall in love with the technology; fall in love with the problems it solves. During the half-day workshop for startups, other speakers echoed the sentiment around engineers that fall in love with their robotics innovation.
“One of the biggest defects or failure modes that I see at startup companies, especially in an audience like this or the startup companies that find their way to my doorstep is a scientist or an engineer that has become emotionally attached to their idea and doesn’t let it go,” said Warren Katz, who works with Techstars and their accelerator program.
#3) The people who use the robots should be more involved in their development
This was an argument made by Howie Choset during his RoboBusiness 2019 keynote. He felt it was especially relevant to the manufacturing industry. “I’d very much like to see the worker more involved in the design cycle of automation,” he said. “Right now, the way it works is this: we make a system, and then we give it to people to use it, and that’s it. But I really think that worker — that end user — is the person who’s going to figure out the best use of these robots.” This isn’t just about empowering the work, he emphasized, but improving productivity and tapping a rarely used source of innovation. “There’s a lot of talent that we’re not using,” he said.
In the long run, demand will increase for non-academic robotics specialists.
This was an idea that emerged from the State of the Industry panel. York Exponential CEO John McElligott talked about creating trade-school programs for “robot mechanics.” As robots move further and further into the mainstream, they will become easier to take care of, and maintenance tasks could be handed off to “the mechanically minded” workers who don’t need a Ph.D. in robotics or computer science. He predicted the emergence of a kind of ecosystem of supporting occupations. His organization is already contributing to this ecosystem, he said. “We’ve got guys that were working at Pier One, and now they’re working alongside robotics engineers,” McElligott said.
#4) Reaching potential customers is a challenge for engineers
Sales and marketing aren’t necessarily natural activities for engineers, but you can’t build a robotics company from the lab. This was the consensus admonition of the “Building a Robot Company from the Ground Up” panel. Though it might seem obvious, this bit of wisdom is often ignored far too long by early-stage startups focused on building a minimum viable product (MVP).
“At the heart of every great product is a customer,” said Josh Lessing, co-founder and CEO of Root AI. Sitting down with the customer and a CAD rendering of the shell of a robot and getting real-world feedback is the right starting point, Lessing said. “I’ve done this many times,” he said, “and I’ve received some pretty harsh and candid feedback — this component makes no sense, I don’t need this because it’s already installed in my facility. Working as just elite technologists starting with a customer demand and working backwards to the engineering that solves their problems; that’s probably the most capital efficient way of actually, meaningfully getting to market.”
“I understand the focus on an MVP,” said Erik Nieves, founder and CEO of Plus One Robotics. “But I think it’s more useful to focus on the MVC, the Minimum Viable Capability, that then can express itself in multiple products. … Yes, it’s important to constrain the problem, so that you can make progress, but don’t get so hung up on ‘this is the product.’ The user will define the product for you. You need to have an engineering wherewithal to build a capability that will manifest itself in a number of different products.”
#5) The robotics industry needs universal interfaces
This idea came up in a number of presentations in different forms. Choset touched on it during his conference-opening keynote, noting that it’s not possible right now to “mix and match” robots in manufacturing settings. We’re now in a period that’s similar to where we were with PCs a decade or two ago, he said. The advent of standards in that industry facilitated “incredible product improvements.”
Dr. Alan Federman, research associate at Stanford University, talked about the value of open source in robotics, and pointed to Stanford’s development of an open source platform for the Vision and Learning Lab (SVL) with the custom-designed social mobile robot “JackRabbot.” The three-year project aims to develop robots that can interact “politely” in chaotic human environments, such as shopping malls, and ultimately develop artificial intelligence models that enable robots to perform this “social navigation” while accomplishing helpful tasks.
The ROS (Robot Operating System), a set of software libraries and tools for building robot applications that started at Stanford, is open, Federman acknowledged, but the industry needs some open-source hardware.
Camilo Buscaron, principal open source technologist at AWS, talked about the future of open-source software in the development of frameworks for robot development. “Typically, robotic companies own the full stack of development for a robot application,” Buscaron said, “which, if you were to tell that to a mobile app developer, or a cloud developer, that would seem insane, because they don’t have to worry about all that. They just have very stable APIs that they can use to build and release their applications. This is why new paradigms are emerging [for robotic software development] that provide for rapid prototyping with software blocks.”
#6) The challenge of getting from “5 to 500” robots
While startup companies in the robotics space have been plentiful in recent years, there’s also a lot of evidence of companies that didn’t quite make it, including some high-profile ones. For many companies that have achieved some initial success with their systems, the next big step for them is to “scale up” and get from those early deployments to mass adoption in the industry that their robot is addressing.
We’re now seeing some new companies that want to help those companies along the way by tackling some of the software, management and monitoring tasks. At RoboBusiness 2019, we spent some time with inOrbit and Formant, two companies looking to provide cloud-based software and services for companies looking to get a better handle on their robotics fleets. Instead of a startup trying to “build something from scratch”, they’ll be able to adopt services like these to deploy faster. See also the offerings from Amazon and its RoboMaker platform, and Qualcomm and its RB3 platform, although those are intended for companies looking to get robots up and running faster.
Additional material contributed by Keith Shaw.