Eight years ago, Morgan Quigley, Eric Berger and Andrew Ng published a paper that was not about ROS. It was about STAIR, the STanford Artificial Intelligence Robot, which used a library called Switchyard to pass messages between software modules to perform complex manipulation tasks like stapler grasping. Switchyard was a purpose-built framework that was designed to be modular and robot-independent, and it was such a good idea that in 2009, “ROS: An Open-Source Robot Operating System” was presented at the IEEE International Conference on Robotics and Automation (ICRA) in Japan. As of this month, the paper introducing ROS has been cited 2,020 times, an increase of more than 50% over last year.

The popularity of one single paper is only a minor indicator of the popularity of the robot operating system that it introduced. At eight years old, ROS is growing faster than ever, and helping the robotics community to grow along with it. We’re especially excited to see how brand new startups have been taking advantage of the open source nature of ROS to help them develop useful, reliable robots that are creating entirely new markets. In 2015 alone, more than $150 million in VC funding (that we know of) was invested in businesses that use ROS.

Large, established companies have been taking more and more notice of ROS as well. At ROSCon this year, Fetch Robotics was joined as a platinum sponsor by Ubuntu, and a record number of gold sponsors included NVIDIA, Bosch, and Qualcomm and attendees from companies such as BMW, DJI, Intel and more. ROSCon 2015 was by far the largest conference we’ve ever had: it sold out weeks in advance. Clearly next year we’re going to have to find a much bigger venue to make room for more attendees, more speakers, and more exhibitors.

Taking a look at how much our community has grown this year, it’s easy to see why ROSCon has become so popular: it’s a reflection of the enthusiasm and engagement of the ROS user base. In May 2015 alone, nearly nine million ROS packages were downloaded from over 70,000 unique IP addresses, and these numbers don’t even count mirrors. This suggests that ROS probably has hundreds of thousands of active users. We also have a very robust developer community: 1,840 people have contributed to ROS’ 10 million lines of code, averaging 20 commits per day. The ROS wiki has gotten 10% bigger since last year, and there are over 11,000 users on ROS Answers, a 32% increase over last year, with a total of more than 5,000 questions answered. It’s numbers like these that make us so confident in the long term future of ROS.

Counting ROS

Because of the nature of the ROS license, we actually don’t know how many users, robots, and developers there are utilizing ROS. Many of the numbers that we are citing throughout are likely to be much larger. For example, we specifically know of approximately 80 types of robots using ROS, but almost every day we hear about new ones. And not every company using ROS discloses so publicly, so our estimates on venture capital investment can be better characterized as lower bounds than estimates.

If you’re not part of the ROS community yet, there’s never been a better time to get involved. Even if you don’t have experience with robots or programming, there’s a wide variety of low-cost robots and helpful online tutorials that can get you started, and we’re also delighted to announce (just in time for the holidays!) that O’Reilly Media has published “Programming Robots with ROS: A Practical Introduction to the Robot Operating System,” by Morgan Quigley, Brian Gerkey, and Bill Smart, which will take you from zero to ROS expert in just 448 pages.

Learning ROS will allow you to do all kinds of cool stuff with more than 80 robotic platforms. You can choose from the capable, affordable TurtleBot, one of the many sophisticated humanoid robots that competed in the DARPA Robotics Challenge, or even NASA’s Robonaut, currently undergoing testing on the International Space Station. Robots powered by ROS are everywhere, and here are just a few of them:

For full frame clips see the long version of the montage.

Of course, we have no idea how many robots are actually running ROS, or how many people or companies are using it, because ROS is open source and completely free. We’re often surprised to learn that cutting edge robots that we’re already familiar with are powered by ROS, as when BMW announced at ROSCon that they’ve been using ROS in their autonomous cars for the past few years. We weren’t at all surprised to hear why BMW chose ROS for its autonomous driving research though: they appreciate its popularity, its stability and reliability based on a large user base, the fact that it makes it easy to collaborate, and its open source nature.

As the ROS community has grown, various special interest groups have organized to promote ROS for specific application domains. The ROS Industrial Consortium is one such group. ROS-I is a software library that builds on ROS and leverages its power and flexibility to control manufacturing automation equipment including industrial robot arms. It is supported by the 36-member organizations comprised of companies such as 3M, ABB, Boeing, Ford, Boeing, Siemens and more. Representatives from Boeing, Caterpillar, Yaskawa and more speak on behalf of ROS and ROS-I in this video.

2016 is poised to be the biggest year ever for ROS, and we’d like to highlight two things that are worth getting particularly excited about. The first is ROS 2.0, which we’ve been developing for the past few years. ROS 2 will support the growth of the ROS community by making it much easier to work with small embedded systems, teams of multiple robots, and robots that require real-time control. We’d also like to make sure you’re familiar with Robotics Fast Track (RFT), which is  a program that we’re working on with DARPA. It’s an easy way for you to get government funding for your awesome robotics ideas without having to give up any of your IP, and absolutely anyone can apply.

We write these anniversary posts to help give you a sense of how ROS has a whole has been doing over the past year, but we’d certainly encourage you to find out for yourself, by getting involved. Write or edit a Wiki page. Answer a question on ROS Answers. Come to ROSCon. And, when you’re ready, think about helping to maintain ROS itself, or even contributing a brand new ROS package. OSRF is doing great, but the long-term success of ROS depends on all of the incredibly awesome ROS users themselves. If you’re already an active part of the ROS community, we can’t thank you enough, and if you’re not, think about it: you can help ROS grow and thrive for eight more years, and beyond.

While arguing by analogy is fraught with peril, I believe that the similarities between robotics and the rest of the IT world are strong enough to justify it.  In robotics, we have many shared problems to solve when developing a product or service, from low-level drivers to high-level capabilities, and all the developer libraries and tools in between.  I have yet to see a successful robotics business for whom any of that stuff is the competitive advantage.  Rather, success comes from the innovative composition and application of that technology in a form that somebody will pay for.  The hard part is figuring out what the robot should *do*.  By working together on the common underlying problems, we end up with better, more reliable solutions, and we free ourselves to spend more time at the application level, which is where we can differentiate ourselves.

In other words, I believe that open source is a great model for the robotics business as a whole.  Now, is it a good model for any individual company?  It certainly can be.  As examples, we see small-to-medium companies, such as Clearpath Robotics, Rethink Robotics, and Yujin Robot, which use ROS directly in their products. And we see larger companies, such as Bosch and Toyota, using ROS in R&D and prototyping efforts.  These are all profit-motivated companies making what is presumably a rational economic decision to rely on open source software.  They’re each holding something back that is their “special sauce,” whether that’s higher level application software, configuration data, customizations to the open source code, or the designs for the hardware.  And that’s expected: unless you’re in a pure consulting business (selling your time), then you need to own and control something that forms the basis of your product or service offering (to allow you to sell something other than your time).

Fortunately, open source software is entirely compatible with such business models.  In fact, it was our hope to one day see such commercial users of ROS that led us to choose a permissive license (BSD, or Apache 2) for the code that we developed.  We’re now witnessing, with the debut of so many new robotics companies, the fruits of those earlier labors in building a shared development platform.

Read more answers →

I say “applications” to distinguish this work from that of developing new algorithms or core building blocks.  Making an analogy to traditional software development, I don’t need to understand how process schedulers, or file systems, or memory managers work in order to develop useful desktop applications.  And I don’t need to know the details of DNS, web servers, or web sockets to develop portable web applications.  Knowing more about the underpinnings of the system will always be useful, of course.  But the key is that, once the building blocks are established, understood, documented, and tutorialized, the barrier has been greatly lowered: you just need to be able to write code.

Beyond just getting more people working with robots, we need better ideas for how robotics technology can be usefully and profitably employed to support people in their everyday lives.  My experience in the robotics community over the last 15 years has convinced me that roboticists are pathologically bad at coming up with application ideas.  We’re enamored of the technology, which is good in that it motivates us to work hard on important problems.  But it also leads us to concentrate on “robotic” solutions to problems, without regard to what people who experience those problems really need.  We can fix this problem by adding orders of magnitude more developers to our community, each of whom comes with a new and different perspective. And we can do that by making the development of robotics applications accessible to any competent programmer.

The Android and iOS platforms made it possible for people with no more than a passing understanding of 3G, GPS, or touch screens to build useful, even world-changing mobile applications.  We can do the same for robotics.  We’re on the right path, with a lot of effort going into open, shared software platforms for robotics.  We just need to keep pushing, and to keep the non-robotics engineer in mind when we’re building things.

Read more answers →