An Eclipse Automotive Stack for rapid-prototyping Software Defined Vehicles
Presented by Sven Erik Jeroschewski (Robert Bosch GmbH) at EclipseCon 2022.
Processing data in a vehicle and getting it out in an uniform way is beneficial for many use cases on the journey to a Software Defined Vehicle (SDV). However, many applications are specific to the underlying hard- and software systems making it difficult to build up a joint stack that can be adopted and extended by solution developers without taking too much effort into the underlying layers. It is desirable to deploy the respective stack and a possible backend not limited to actual vehicles and cloud backends but also to locations more accessible to developers at their desks. Such a portable stack enables a quick start and fast feedback loops during the development and makes it easier to get used to this technology.
A rapid prototyping stack requires a hardware platform, an operating system, a run-time environment, and tooling. In the presented setup, we use the Eclipse Kuksa CANOPi as a hardware platform around the Raspberry Pi Compute Module 4. For the operating system and the run-time environment, our demo setup relies on an Eclipse Leda based image and Eclipse Kanto to manage the execution of containers on the vehicle. In addition, Eclipse Kanto allows the retrieval of data from Kuksa.Val based on the Vehicle Signal Specification (VSS) and then send this data to a backend application. In addition, Eclipse Velocitas can be used to accelerate the development of specific vehicle applications.
With this talk, we share our findings and learnings of setting up a technology stack for an SDV by deploying software and hardware components from the Eclipse Automotive top-level project.We intend this talk for other open-source developers and technology teams looking for a base technology stack to build their SDV applications or options to add in more technologies.