The talk will show how the Eclipse KUKSA project can help to leverage the advantages of a standardized description of signals based on VSS inside a vehicle.
More and more computing power is available in modern cars. This development has tow main drivers: Firstly modern Autonomous/Assisted Driving functions requrie raw computing power, secondly more and more of a vehicle's value is realised in software that is expected to be updated during the lifetime of a vehicle and be applicable across a wider range of vehicles. This leads to systems that look more "IT like", where you have processors, GBytes of RAM, Hypervisors and maybe even Linux cells and containers. At the same time security requirements and providing or coexisting with safety requirements in a vehicle need specialized solutions
A large challenge for automotive software is the lack of standardization inside a vehicle. In large part the basic data fabric of a vehicle consists of "signals", such as for example the speed or temperature of a compenent. For this kind of data traditionally there is no standardized semantic and no standardized way how it is represented or transmitted. This makes even seemingly simple end-to-end applications such as "upload mileage to the cloud" very hard in the industry, even when only limited to a single vendor. The COVESA Vehicle Signal Specification (https://covesa.github.io/vehicle_signal_specification/) tries to address this by introducing a domain taxonomy for vehicle signals. The goal is to create a common understanding of vehicle signals in order to reach a “common language” for vehicle data independent of the protocol or serialization format.
The talk will show the archiectural answer given by the Eclipse KUKSA project to leverage such a standardized signal description inside a vehicle: Running on a vehicle computer, we will show how to use KUKSA components to gather and transform the data from lower embedded layers in the vehicle, and how to provide access to apps and services using that data. We will discuss common use cases such as using the data for an in-vehicle functionality, transmitting vehicle data into the cloud as well signal refinement, whereas lower level signals are converted to higher level ones. We will take a look the APIs provided in KUKSA and touch upon questions of data security and privacy. This should act as a example why within the context of SDV, Vehicle Computers are a worthwhile target for OSS solutions, where a lot of benefit can be created by adaopting standaridzed solutions for the common challenges, laying a solid and cost-effective foundation for all kind of applications and services on top.