Robotics

QERRA-v2 Classical brings explainable ethical scoring to ROS 2 robotics

A fully traceable ethical AI using 12 human-centric vectors, no black boxes.

Deep Dive

QERRA-v2 Classical is a 100% classical, fully explainable ethical scoring engine built around 12 immutable human-centered vectors called SEMEV-12. Unlike most AI safety tools that rely on opaque neural networks, this system returns a traceable score and step-by-step reasoning, making every ethical decision auditable. The engine runs standalone or as a ROS 2 node, subscribing to /qerra/situation_input and publishing decisions, confidence scores, and the complete SEMEV-12 result on separate topics.

Designed for humanoid robots and autonomous systems, the framework is open for community feedback — particularly on how the topic structure and message types fit existing robotics pipelines and whether the bridge could be more useful in real Behavior Trees. A live API and full documentation accompany the release. This is the latest iteration of the QERRA project, which previously explored hybrid quantum-classical approaches for ethical safety layers.

Key Points
  • Based on 12 immutable human-centered vectors (SEMEV-12) for transparent ethical scoring.
  • Fully explainable with no neural networks — returns traceable scores and reasoning.
  • Includes a ROS 2 bridge with subscription to /qerra/situation_input and publishing of score, decision, and full SEMEV-12 result.

Why It Matters

Brings transparent, auditable ethics to autonomous robots — critical for trust and safety in real-world deployments.