Robotic Operating System (ROS) is a middleware SW layer and toolset for the creation of robotics solutions. In the Autonomous Mobile Machine Research Group from Tampere University, led by Associate Professor Reza Ghabcheloo, we are actively using ROS in the heavy machine automation research work. Our interest to ROS is twofold: the platform itself is subject of the study in PEAMS project, but also ROS is used actively in other research activities, as well as in education.
Research on SW Platforms for Heavy Machines
The research in the PEAMS project aims to understand the benefits of using common SW-platform and related ecosystem for the heavy machine industry. The central part of the work is to create a reference architecture and experiment with that in real machines. Besides the technical aspects we also do research how industry is currently seeing the use of SW platforms and how this expected to evolve in the near future. The hypothesis is that common platforms and access to open-source software ecosystems, will increase the R&D efficiency, and collaboration between companies as well as research institutes. On the other hand, adopting the new architectures and technologies is never a simple task.
As a platform our specific research interest is ROS2. It is currently reaching the level of maturity and stability that makes it viable solution for real life products. While previous versions of the ROS have already been utilized in commercial products, the platform has mainly been used by the research community and with smaller companies. But now it is seen that ROS2 is gaining more wider attention. For example, a recent article in Science Robotics, “Robot Operating System 2: Design, architecture, and uses in the wild” by Steven Macenski et al. gives nice recent review of the ROS2 with some usage examples in different application domains.
ROS2 as Tool for The Research
The research in mobile robotics is a challenging multidisciplinary field. Our research group focuses on innovative solutions, methods, and tools to build autonomous applications. These include topics like, autonomous navigation and manipulation, sensing and robot learning, and optimal control, to name a few. The primary interest is with advanced approaches that solve real-life problems on solid theoretical grounding.
In practical level such research requires collaboration between researchers of various technical backgrounds and creation of high complexity experimental HW and SW platforms. The work needs to also easily transferrable between different machines and simulation environments.
ROS platform and ecosystem are supporting this kind of work very well. The modular SW architecture guides the creation of good quality maintainable SW that is equally important in long term continuous research work as in product development in the industry. Also, the use of solutions and tools provided by ROS community accelerates the work as there is no need to “re-invent the wheel” for common functionality, but researchers can directly focus on studying and developing novel algorithms and solutions.
We are also already observing that ROS2 is helping to facilitate co-operation between academic world and industry. In joint applied research projects, where the focus is to gain experience how latest research results could be utilized in the industry, ROS2 offers a nice foundation for modular architectures that allow easy and fast integration of components from different parties together.
We are looking forward discussing about our experiences further via the PEAMS project seminars and invite industry and other research groups to share their viewpoints and insights also in this forum.
Written by
Jukka Yrjänäinen
Tampere University