Architecting ADAS/AD Applications: Why YAML is a Good Choice

15th April 2025

Advanced Driver Assistance Systems (ADAS) and Automated Driving (AD) applications are characterized by complex system and software architectures and strict safety requirements. When describing these systems, using the right data representation language is crucial — not just for technical accuracy, but also for maintainability, readability, and collaboration across teams.

For , we deliberately chose YAML as the foundation for the architecture specification language YAAA. YAML offers the best balance between human readability, expressiveness, and machine-parsability, making it an ideal choice for describing modern, safety-critical automotive systems.

Challenges in Describing ADAS/AD Architectures

ADAS and AD systems integrate multiple, highly specialized components:

• Sensors and sensor fusion
• Perception and environment modeling
• Path planning and control
• Communication between distributed computing units

These systems must meet strict safety (e.g., ISO 26262) and real-time performance requirements. Therefore, a description format must be:

• Easy to read and verify by both developers and safety engineers
• Flexible enough to model a wide variety of configurations
• Scalable to handle growing system complexity
• Robust and structured for machine processing

Data Representation Formats: A Comparison

Before deciding on YAML, we carefully considered other widely-used formats. The table below highlights key differences:

Why YAML Was Chosen for ETAS DMS

Given the above, YAML clearly stands out as the best fit for the goals of ETAS DMS:

• Clarity: YAML is highly readable even in complex models, making it easier for teams to validate architectural descriptions manually, which is important in safety-critical workflows.
• Flexibility: YAML handles complex nested structures and supports anchors and aliases, enabling modular and scalable system descriptions.
• Developer Productivity: Writing and maintaining YAML is less error-prone compared to XML or JSON. This accelerates development cycles and simplifies onboarding new team members.
• Modern and State-of-the-art: YAML has become the de-facto standard for configuration and infrastructure description in many domains (e.g., Kubernetes, GitHub Actions). It reflects modern best practices for clean and maintainable system design.

By basing the ETAS DMS DSL on YAML, we provide a solid, future-proof foundation for modeling even the most complex ADAS/AD architectures—one that meets the needs of today’s developers, engineers, and safety teams.

Learn More: YAAA - YAML As Architecture

If you’re curious how YAML is practically used to model complex ADAS and AD systems, we invite you to explore YAAA (YAML As Architecture) — the domain-specific language behind the EDMS solution.

YAAA provides a structured way to describe functional components, communication interfaces, and deployment information for automated driving systems, using clear YAML-based templates.