By Prof. Wyatt Newman, Ph.D.


As a researcher and educator in robotics, I’ve been utilizing and following the Robotic Working System (ROS) since its early days, greater than 10 years in the past. With the current introduction of ROS 2, the system has reached a serious turning level. Extra than simply an improve from the unique, this new model is an entire rethinking of ROS—a real evolution that delivers a strong ‘Act 2’ for a framework that has already been a key accelerator for the sphere of robotics. The brand new model contains key variations and vital incompatibilities—each of which demand the eye of the robotics group and a renewed give attention to contributing to this very important enabler of the way forward for robotics.

Two years in the past, I wrote the ROBO World article ROS is Fueling In the present day’s Robotics Renaissance through which I highlighted the significance of the ROS open software program ecosystem and its function within the development of robotics. What I surmised on the time—that ROS would proceed to play a key function within the resurgence of robotics—has come to fruition. ROS has developed quickly during the last decade. The end result has been 13 main releases since March 2010 and a system that has dramatically diminished the programming barrier for roboticists across the globe. By serving to programmers overcome the ‘Tower of Babel curse’ with a single, open-source language and the power to reuse code, ROS has dramatically accelerated the tempo of innovation and buoyed robotics startups that might not have succeeded with out the system’s re-usable packages. In the present day, ROS has grow to be the de facto world normal for robotics, with tons of of 1000’s of roboticists utilizing and contributing to the ROS ecosystem.

Clearly, I’m an enormous fan of ROS. So when it was introduced that the Could 2020 launch of ROS can be the final to be supported by the Open Supply Robotics Basis—and that every one future releases would assist ROS 2—I used to be admittedly shocked by this main shift in technique. Greater than a easy rebranding of the present ROS, ROS 2 is totally different in quite a few basic methods, and it contains necessary incompatibilities with ROS that require an entire rethinking of the system. These adjustments introduce a wholly new department within the evolution of the ROS system, in addition to a brand new course for the sphere of robotics itself. Given the large increase ROS has dropped at robotics via its capability to standardize on ROS, this new course at first appeared to be an odd selection. Nevertheless, after diving deeper into the brand new framework, I’m optimistic that ROS 2 will assist robotics advance to even higher ranges.

5263-Wyatt-Article-Image-(V001)_A (1)

One motive the unique ROS had remained viable for so long as it did was that the founders of the system possessed the foresight to acknowledge that robotics can be an evolutionary course of. With that in thoughts, they constructed a system that exceeded expectations from the beginning. Regardless of this forward-thinking strategy, sure facets of ROS have been strained to the restrict as robotics developed over time. Foremost amongst these limitations was the underlying community communications middleware. The unique authors created their very own communications mechanisms from scratch, and, to their credit score, these held up remarkably properly for years, whilst distributed pc techniques grew to become more and more advanced. However there have been sure facets of that evolution that proved inconceivable to foretell. Listed below are just some examples:

Swarm robotics

Within the unique system design, communications required a single ‘ROS Grasp,’ and all computer systems within the distributed system required a secure connection to that central hub. The event of swarm robotics revealed this construction to be a weak point. In a swarm of drones, for instance, communication between the system and the person drones is commonly sporadically damaged after which rapidly re-established—a situation that isn’t supported by the unique ROS.

Web of Issues (IoT)

IoT has delivered a large inflow of low-cost and ubiquitous gadgets. Greater than 100 new IoT gadgets are being added each second and, like swarm robotics, these gadgets be part of and go away networks unpredictably—a actuality that isn’t appropriate with ROS networking assumptions.


Safety in robotics is significant, and the dearth of robust safety protocols within the unique ROS has been a rising concern for roboticists. This has introduced a barrier to adoption in trade and posed a really actual risk to enterprises and shoppers which have grow to be more and more depending on robotics.


During the last decade, web communications ‘middleware” has superior tremendously. Middleware is the important thing to managing complexity and making certain high quality of robotics techniques. It helps integration, simplifies software program design, and permits communication between sensors. Over time, nonetheless, the middleware that was inbuilt to the inspiration of ROS has grow to be outdated and unable to assist the rising variety of new use instances.

5263-Wyatt-Article-Image-(V001)_D (1)

For these causes and extra, it was essential to make vital adjustments to the foundational elements of ROS. ROS 2 is the answer. The brand new system is able to operating on small, embedded platforms, making it properly suited to trip the wave of IoT by eliminating the requirement to attach new gadgets and gadgets to a grasp hub; as a substitute, gadgets will be found, related, and re-connected independently. Different necessary issues have been additionally built-in into the brand new framework. Whereas ROS was restricted to Linux-based techniques, ROS 2 is appropriate with Home windows. ROS 2 additionally incorporates know-how that helps actual time processing—roughly to the 1 millisecond stage—which can be a serious profit to interactive techniques, security techniques, and motor controls.

It is very important word that robots constructed on the unique ROS are usually not being deserted—no less than not but. Although damaged APIs are inevitable when migrating to ROS 2, a ‘bridge’ was designed to permit for gradual, incremental upgrades to the brand new framework and to assist heterogeneous techniques of ROS and ROS 2-powered robots. However whereas wholesale upgrades aren’t instantly crucial, it’s clear that the way forward for ROS is ROS 2. And trade is already recognizing the significance of the change. Simply this month, Qualcomm introduced its RB5 robotics platform that integrates 5G communications with AI, machine studying, machine imaginative and prescient, and ROS 2. Adlink has additionally jumped onboard, collaborating with Intel to launch the ROScube-I ROS 2 Controller in an effort to “notice AI Robotics on the edge.”

Nobody expects the transition to ROS 2 to be painless. That mentioned, there may be little doubt that ROS 2 is the important thing to reaching even higher developments and capabilities in robotics. Programmers, industries, shoppers, and buyers are positive to reap the advantages.



Prof. Wyatt Newman, Ph.D. has held visiting appointments at Sandia Labs, NASA, and Princeton, and worldwide appointments at Philips (Eindhoven, The Netherlands), as a Distinguished Visiting Fellow at U. Edinburgh, and The Hung Hing Ying Distinguished Visiting Professor at U. of Hong Kong. He has led groups in two DARPA grand challengesautonomous automobiles and catastrophe response robotics—and in two current NIST competitions for agile robotics. His textbook, A Systematic Method to Studying Robotic Programming with ROS, was launched by CRC Press in 2017.