• Johan Ekberg

AUTOSAR and ROS 2 for Software-Defined Vehicle

The ROS 2 community and the AUTOSAR consortium are both addressing the challenges of creating a middleware that can be used in complex embedded systems. ROS 2 has taken a code-first approach to make it as easy as possible to develop new applications for such systems. Developing with ROS 2 quickly leads to a first working product in modern C++ code that over iterations becomes stable production software. AUTOSAR has a strong focus and successful history in creating interoperability between suppliers and OEMs by providing specifications on how to behave on vehicular networks. AUTOSAR has created a structured exchange format that makes it possible to share requirement-related information like interfaces and parameters between organizations.

Figure 1: AUTOSAR and ROS are better together. Source: Apex.AI

Apex.AI is working to bring these two communities together by creating a middleware that is both developer friendly and supports the interoperability with standards provided by AUTOSAR.

ROS 2's design and foundations make it perfectly suitable for the automotive domain; however, improvements and extensions are needed. Therefore, Apex.AI has improved and extended ROS 2 to support the requirements for hard real-time systems and functional safety. These improvements are implemented in Apex.OS and Apex.Middleware.

Apex.OS is certified by TÜV Nord as a Safety Element out of Context (SEooC) up to ASIL D according to the ISO 26262 standard, and Apex.Middleware is planned to be certified to the same safety level in 2023.

Apex.Middleware is a stand-alone product that is the officially supported communication middleware for Apex.OS. It is based on Eclipse iceoryx™ and Eclipse Cyclone DDS™, and it can be used for both inter-SOC and inter-ECU communication. Apex.AI has developed connectors for communication protocols, including SOME/IP, CAN, and MQTT to further extend the communication capabilities.

The SOME/IP and CAN connectors connect the ROS 2 community with the rest of the vehicle communication network. This gives the automotive developers a rich set of community-contributed ROS 2 packages for sensors, computer vision, point clouds, actuators, data recording/playback, and much more.

Figure 2: AUTOSAR and ROS are unified in Apex.AI products, Source: Apex.AI

The picture above shows how the SOME/IP and CAN bridge provided by Apex.AI connects an Apex.OS Node with both the AUTOSAR Adaptive and the AUTOSAR Classic platform. On the left side of the picture, it is shown how AUTOSAR arxml, containing descriptions of data-types and interfaces, together with the DDS IDL definition can be used as configuration input to the bridges. This makes the setup flexible for changes in the service definitions. In the middle, it can be seen how published and subscribed topics in the Apex.OS node are converted to AUTOSAR types (by the data and signal conversion layer) and then published to the networks. The conversion routine can either use a routine provided by the developer or a default routine provided by Apex.AI. The converted messages are then sent to (or received from) the networks and can then be used in AUTOSAR Adaptive applications using ara::com or in AUTOSAR Classic RTE.

Apex.AI has its foundation in the ROS 2 community and contributes actively to new releases of the open-source project. In April 2020, Apex.AI joined the AUTOSAR consortium with the clear goal to bring the AUTOSAR ecosystem and ROS 2 community together. In the 2021 ROS 2 Galactic release, a number of new features were contributed by Apex.AI and you can read about them here. In May 2022, the ROS 2 Humble Release includes Apex.AI contributed Eclipse iceoryx 2.0 Blueberry which is functionally complete for AUTOSAR ara::com zero-copy interprocess communications to supercharge your Adaptive AUTOSAR applications.

At the AUTOSAR Open Conference 2022, Apex.AI CEO Jan Becker presented the driving forces and benefits that are making the best of these two worlds come together.

Request a product demo to see how your team can live in the AUTOSAR and ROS 2 worlds: info@apex.ai.

If you are interested in Apex.AI products for your projects, Contact Us

We are also growing our team worldwide. Remote and in our Labs. Visit Careers