Which IDE is best for ROS

Who is ROS for?

As open source software, anyone can use ROS free of charge. Since it is a modular framework with functions for sensor processing, evaluation, planning and control, ROS is not only interesting for robotics, but for many technical systems. In the specialist field for autonomous driving, at least parts of ROS are also often used. In robotics, almost all research groups now use at least some of the ROS. Many research groups no longer have their own software framework, but concentrate fully on ROS. Or they combine ROS with their own framework, such as the FZI.

  1. German Climate Computing Center (DKRZ), Hamburg
  2. DRÄXLMAIER Group, Vilsbiburg near Landshut

But there are also some large companies that use ROS intensively in their research departments. Often this also happens together with a research institution. In addition to large companies such as Bosch, BMW or Schunk, many smaller companies and startups also use ROS. Companies and research institutions have joined forces in the ROS Industrial Initiative to make ROS even more robust and suitable for industry.

ROS cannot (yet) do that

ROS can make the development of new applications or algorithms easier in many places. But it does require some preliminary qualifications. ROS is a software framework in C ++, Python and Java. Programming skills in these languages ​​are therefore required.

In addition, there is only very rudimentary support for Windows so far, so a Linux development environment should be used. In addition to this general programming knowledge, experience with the numerous ROS packages is also important. One disadvantage of open source also affects ROS: not all packages are of the same quality or are still maintained by their developers. Choosing an outdated or poorly performing ROS package can quickly become very frustrating and discourage beginners in particular.

Real time - does not work

The communication structure of ROS does not yet support real-time execution. This makes it difficult to use in safety-critical areas. ROS is also a central system with a master. In many applications this is not a limitation, but in very decentralized approaches it can be limiting.

Overall, the software standard is above average, as there is a peer review process before new parts of the code are added. However, ROS has not yet been certified. Anyone who wants to use the advantages of ROS therefore still has to bring a lot of expertise with them. Even with ROS, new robot applications do not simply fall from the sky.

What's next for ROS?

ROS has united robotics research, significantly improved the collaboration and reusability of software and is currently slowly fighting its way into commercial, industrial applications.

Even though Willow Garage pulled out a while ago, ROS development continues unabated. The community is tackling various weaknesses of ROS piece by piece. In the meantime, for example, several masters can also be operated in ROS at the same time. The topic of real time, which is important for industry, can be reached in a few detours - e.g. B. ROS nodlets and lots of CPU power. In addition, the community is working on ROS 2.0 with a completely new level of communication that can then also support real-time in the future. Like numerous ROS developers, the FZI is committed to the professionalization of ROS.

Fanuc industrial robot at Automatica 2016 with ROS (Image: FZI)

The ROS Industrial Consortium (ROS-I) works together with well-known international companies to make ROS fit for industry. With success - the interest of companies is increasing. More and more providers of robotic hardware are delivering their systems directly with ROS drivers or interfaces.

Arne Rönnau, Head of the FZI Living Lab Service Robotics, relies on ROS as the basis for the development and rapid commissioning of new robot applications in numerous projects. The FZI uses ROS for its own robot systems such as Hollie, Lauron V or Bratwurst-Bot, but also supports companies in using the open source framework.

  2. 1
  3. 2
  4. 3
  5. 4