The Eclipse SDV Hackathon

The automotive industry is undergoing a transformation towards Software-Defined Vehicles (SDVs). This development allows for continuous improvement through software-based updates, personalization, data-driven services, and the integration of new features throughout the vehicle's lifecycle. The Eclipse SDV Working Group hosts an annual hackathon to advance open-source projects in the context of SDVs.

The Eclipse SDV Hackathon 2025 provided us with the perfect opportunity to apply the SDV ecosystem in a practical scenario – with great success! Our CarByte team took 1st place. This demonstrates that we are successfully navigating the SDV ecosystem and mastering the diverse open-source solutions. Additionally, we are engaging in the open-source community as the code from the hackathon is intended to be made publicly available as the official Eclipse SDV Blueprint.

Open Source Based Ecosystem for Developing ADAS Functions

In this year's hackathon, we implemented an ADAS function. To do this, we utilized open-source technologies such as KUKSA, VSS, Ankaios, and Podman in combination with modern communication protocols like Zenoh and uProtocol.

The implemented ADAS function allows a truck driver to interact with their vehicle. The interface with the driver consists of an agentic AI app featuring Speech-to-Text and Text-to-Speech. In addition to the general knowledge accessible to an LLM, the assistant is equipped with knowledge about the vehicle. This is achieved through the use of KUKSA, which provides access to vehicle data.

The framework of the hackathon has defined the landscape of the SDV ecosystem. We decided to implement a practical end-to-end solution that integrates as many SDV projects as possible – similar to those we encounter in everyday life. The implementation consists of three components: the vehicle computer (Compute Node), a physical control unit (MCU Node), and a simulation computer (HPC Node).

Compute Node

The ADAS function operates on the vehicle computer. The function provides a verbal user interface that can output current vehicle information, such as speed or interior temperature. Users can interact with the system via voice input. Data is exchanged in the background using various communication technologies with a physical control unit (MCU Node) or the simulation (HPC Node). For this purpose, uProtocol was used over Zenoh as well as MQTT. The application gains access to standardized vehicle data according to the Vehicle Signal Specification (VSS). The VSS is a hierarchical data model that allows uniform and manufacturer-independent communication between vehicle components, thereby significantly facilitating the development, integration, and reusability of software functions.

MCU Node

ThreadX orchestrates the system. The components were implemented using Rust. This allows us to demonstrate that the ADAS function can access data from a physical control unit.

HPC Node

A driving simulation runs on the HPC Node. Particularly for the first phase of developing ADAS functions, a simulation environment is helpful. Here, vehicle data such as speed, acceleration, and environmental signals are generated. This data is provided via an interface through the KUKSA Databroker for the ADAS function.

Presentation of SDV Technologies

The successful implementation is based on the combination of several open-source technologies:

• KUKSA:  This platform enables an abstraction of vehicle signals for the application in the vehicle, cloud, or peripheral devices. KUKSA serves as middleware that mediates signals from providers to consumers.

  
  

• VSS:  The Vehicle Signal Specification is a standardized data structure that enables consistent signal processing. By abstracting vehicle data, functions can be developed and reused independently of the underlying architecture.

  
  

• Zenoh and uProtocol:  These communication protocols ensure efficient and low-latency transmission of signals between the application and the control unit.

  
  

• Ankaios and Podman:  The containerized architecture allows running various software components in isolated environments. Ankaios manages the containers on a board, while Podman ensures isolated execution.

  
  

• CARLA Simulator: The simulation tool can be used during the development phase to enhance the ADAS application within simulated test scenarios.

What does CarByte deliver?

The first place at the Eclipse SDV Hackathon 2025 underscores CarByte's expertise as a versatile partner for OEMs and Tier-1 suppliers aiming to master the transition to software-defined products. The implemented project provides OEMs and suppliers with a reference for how Eclipse solutions can support the development of new functionalities. With the future publicly available Eclipse SDV Blueprint, we position ourselves as active contributors in the SDV ecosystem and empower customers to implement standardized, scalable solutions for the future.