Microservice Architecture
ROS-package-level modules can be developed independently while CI/CD keeps them integrated in the overall system.
An Open Ecosystem for Collaborative Development and Benchmark-Driven Evaluation
OpenADS explores how distributed partners can build compatible automated driving services and use benchmarks to identify stronger system compositions.
The work is embedded in Germany's Ecosystem Mobility 4.0 initiative and aligned with the European Connected and Autonomous Vehicle Alliance .
What is OpenADS?
An open ecosystem for Automated Driving Systems.
OpenADS is a collaborative, open-source ecosystem for Automated Driving Systems.
- OpenADStack is the modular ROS 2 reference implementation of an automated driving stack.
- OpenADServices are reusable modules for localization, perception, understanding, planning, and control.
- OpenADSim provides closed-loop simulation and scenario-based testing with CARLA or SUMO.
- OpenADSuite provides tools and templates that streamline development workflows.
- OpenADSafety is developing benchmarks and tools for verification and validation.
Containerization
Each module defines its dependencies and container image, so developers can start work without running the full stack.
Flexibility
Modular interfaces support evolving ADS architectures without restructuring the entire system.
Transparency
Open CI tests and benchmarks make contributed modules easier to evaluate, compare, integrate, or remove.
Tutorial focus
Collaborative foundations,
ecosystem applications,
and hands-on interaction.
The tutorial explains why collaborative OpenADS development matters, introduces the reference architecture and real-world applications, and concludes with a 50-minute hands-on session.
Motivation
Why shared interfaces, reusable services, and reproducible evaluation are needed for open AD systems.
Architecture and applications
Relate functional architecture, integration architecture, and deployment composition to data, validation, and fleet-management use cases.
Hands-on session
Take a guided tour through the core OpenADS repositories and then run OpenADSim before exploring scenario-based execution, module benchmarking, and metric calculation with an initial focus on perception.
2.5-hour program
Agenda
Welcome and Introduction
Scope, learning objectives, and setup.
All organizersWhy Collaborative OpenADS Development?
How Ecosystem Mobility 4.0 enables collaborative OpenADS development across research and industry.
Timo Woopen, Thinking Cars GmbHOpenADS Reference Architecture
Functional architecture, integration architecture, and deployment composition: how reusable OpenADServices are assembled into the modular ROS 2 OpenADStack and embedded in vehicle or simulation setups.
Raphael van Kempen, RWTH Aachen UniversityOpenADS in Practice: Ecosystem Applications
Three examples of how OpenADS supports practical use cases.
- Data Foundations for Development and Testing Stefan Orf and Ahmed Aboulazm, FZI Forschungszentrum Informatik
- Benchmark-driven Qualification of OSS-Modules Timo Woopen, Thinking Cars GmbH
- OSS-based Fleet Management Solutions Markus Zarbock, Starwit Technologies GmbH
Lunch Break
OpenADS Hands-On
A guided tour through the core OpenADS repositories and their relationships, followed by running and exploring OpenADSim. Building on this, participants are introduced to scenario-based execution and learn how modules can be benchmarked by calculating metrics, starting with publicly available perception modules and evaluation examples.
Christian Geller, RWTH Aachen UniversityRaphael van Kempen, RWTH Aachen University
Wrap-Up and Summary
Lessons learned and next collaboration steps.
All organizersPractical setup
Bring curiosity and be ready to work with OpenADS.
The tutorial is designed for intensive participant interaction. Motivation, architecture, and ecosystem applications frame the problem before participants work directly with OpenADS.
The 2.5-hour format keeps the tutorial compact and focused: a short setup, a motivation talk, an architecture overview, practical application examples, and a hands-on session.
The hands-on starts with a guided walkthrough of the core OpenADS repositories and then moves to running OpenADSim in a prepared environment. It then introduces scenario-based execution and a first benchmark workflow for calculating metrics on perception modules.
Target audience
Researchers, engineers, suppliers, and industry stakeholders interested in OpenADS, supplier integration, service selection, simulation-based evaluation, and pre-development validation.
Industry value
Faster integration, more transparent software selection, and reduced risk through reproducible cross-partner validation.
Organizing team
Academic, research, and industry partners.
Univ.-Prof. Dr.-Ing. Lutz Eckstein
RWTH Aachen University
Prof. Dr.-Ing. J. Marius Zöllner
FZI Forschungszentrum Informatik
Dr. Timo Woopen
Thinking Cars GmbH
Christian Geller
RWTH Aachen University
Markus Zarbock
Starwit Technologies GmbH
Dr.-Ing. Philip Schörner
FZI Forschungszentrum Informatik
Contact
Questions about OpenADS or the tutorial?
Reach the organizing team at opensource@ika.rwth-aachen.de .