We’ve said it here before, but it’s worth repeating – robots fail, and they fail more often than you think. When companies talk about the benefits of robotics, they tend to gloss over some of the downsides of autonomous systems, especially those that work in dynamic, changing and chaotic environments. Robots get stuck, they get lost, they break down … and occasionally they even catch fire or fall down an escalator.
Robots and Humans are Alike - We All Fail
At InOrbit, our mission is to accelerate the adoption of robotics at scale. After talking to 100+ robotics companies, from startups to companies with billions of dollars in revenue, with all types of autonomous robots, we’ve reached an inescapable conclusion: they will all fail.
We’re not saying the companies will go out of business (although, sadly, many have in recent years). Rather, every robot will experience failures, in some cases multiple times per day or per hour. While a few of these failures may be catastrophic (and occasionally hilarious), like steering into a pond, getting stuck next to a trash can, or suddenly catching on fire, many errors are recoverable. The most frequent failures are what we call autonomy exceptions, where a robot finds itself in a situation that falls just outside its operating parameters.
When I was a child, my father would come home from work and give me the extra computerized punch cards that were used that day - it was my first exposure to the world of computers. Later, I became an active participant in the personal computer revolution, both at school (which let students program on Apple II systems) and at home (my father purchased an early IBM PC system).
Because PCs were so new, most people who used them needed to learn how to program them, and in my case it was figuring out the BASIC language. After a while, my interests changed from programming to writing, but I had friends who stayed interested in programming. They discovered how much more they could do when they programmed on top of libraries that gave them higher functions. Similarly, in the early days of the commercial Internet, I was reading books on HTML so I could create my own web pages. I then also learned how to configure and manage early wireless networks (ah, the joys of 802.11b!).
Technology developments like these often follow the same path - a new technology emerges and gains momentum, but then others begin creating tools that let additional people participate in the process from a different starting point. For example, in the world of the web, you can now build a website by using templates and tools from WordPress or Squarespace, and download thousands of widgets that run different parts of the site, such as creating an online store.
Smarter, autonomous robots, have been developed over the past five years thanks to advances in mobile computing, sensors and AI. Many of these robots are now being deployed to assist in the fight against COVID-19 in an effort to “flatten the curve” of cases or provide human-augmented services for companies providing essential functions.
Over the last two decades, we’ve seen the evolution of hardware virtualization that resulted in the cloud as we know it today. It started with Virtual Machines and then expanded to include Software-Defined Networking (SDN) and Software-Defined Storage (SDS).
Software-defined networking is “an approach to networking that uses software-based controllers or application programming interfaces (APIs) to direct traffic on the network and communicate with the underlying hardware infrastructure.” (Source: VMWare) Similarly, software-defined storage separates the management and provisioning of storage from the underlying physical hardware.
As the COVID-19 epidemic continues to wreak havoc in countries around the world, there’s some reason for hope as the number of new cases in China drops to 0. This shows that with concerted effort by individuals, companies and government, it is possible to limit the exponential scale of the virus.
For this post, I want to go beyond the day-to-day commentary on the robotics industry or our own business and take a look at longer term trends shaping society. Needless to say, this is a personal perspective and pure speculation. Making predictions about what’s to come is easy, getting it right is hard, so take this with a grain of salt. My hope is that this will spur some healthy discussions and drive additional innovations.
The start of a decade is always a good excuse for predictions. After all, who’s going to remember 10 years later if you were right? For us at InOrbit, we are setting a goal and making a prediction that will determine the fate of our company and the broader robotics industry. We fully expect to be held accountable to it.
We recently got together as a team to work on our 2020 Vision (yes, pun fully intended) to guide us through the decade we are just kicking off. As part of this we have committed all our energy and passion behind a BHAG: a big, hairy and audacious goal. If you are not familiar with the concept of BHAG (pronounced “bee hag”), it’s a term popularized by Jim Collins, of Good to Great fame, to capture a simple goal that help align a whole organization.
Last week InOrbit was at ROScon, one of the most important technical conferences in robotics. It was great to see a growing and very engaged community tackle a wide variety of topics. Most of the energy in recent years has been around ROS2, and this year it was clear that it’s really happening. You could find robots on the exhibit floor running ROS2, and most of the talks tackled new capabilities in ROS2.
We noticed another, perhaps more significant change: there was a significant uptick in the number of talks and discussions regarding scalability of operations in the field and interoperability across robots. As ROS adoption is growing, attention is shifting towards managing hundreds or thousands of robots outside the lab.
The opening keynote set the tone. Selina Seah, Director of the Centre for Healthcare Assistive & Robotics Technology (CHART) at Changi General Hospital, and Morgan Quigley, Chief Architect at Open Robotics, presented their ongoing work to coordinate navigation of heterogeneous robotic fleets and more broadly addressing the need for interoperability of complex and disparate technological systems with HIT and infrastructure.
Our very own CTO/co-founder, Julian Cerruti, and one of the engineers on the InOrbit team, Florencia Grosso, presented some lessons learned working with robotics companies to deploy and operate fleets of ROS-based autonomous robots in production. Several other presentations mentioned fleet management, and the excellent panel on ROS at Scale delved into the details of how to handle large systems of robots in production settings.