QERRA-v2 Classical: Fully Explainable Ethical Engine for Robotics Goes Live with 12 Vectors

Marussa Metocharaki, an independent researcher from Greece, has released QERRA-v2 Classical v1.8.8, a fully explainable ethical evaluation engine for robotics and autonomous systems. The engine uses the SEMEV-12 framework, which defines 12 distinct ethical vectors (e.g., beneficence, non-maleficence, autonomy, justice) to score robot actions. Every decision comes with a transparent numeric score and plain-language reasoning — no neural networks, black boxes, or training data involved. The updated public API on Hugging Face now exposes all 12 vectors for developers to call programmatically.

Despite the technical maturity, Metocharaki is now reaching out to the ROS 2 community for integration guidance. Her main question: how to best wrap QERRA as a Condition node inside a Behavior Tree, so that an ethical check runs synchronously before an action node executes. She's also interested in real-world testing patterns, decision-making node best practices, and pitfalls when external ethical checks gate autonomous behavior. The post reflects a rare combination of deep ethical design and pragmatic robotics engineering, and hints at early-stage collaborations (the researcher is also seeking a technical co-founder).