CSF: Contextual Safety Filtering for Motion Generators
What it is
This paper introduces Contextual Safety Filtering (CSF), a training-free method to make text-conditioned motion generators safe by interpreting scene context. It grounds natural language safety rules in reference trajectories and uses a safe reference tracking CBF-QP to enforce constraints on the generated motion.
The problem it solves
Existing safety safeguards for motion generators either ignore scene context (same action can be safe or unsafe depending on what it targets) or require expensive labeled data and specific geometric constraints, failing to account for how context changes motion meaning.
What you could use it for
- Ensuring humanoid robots do not perform dangerous actions around humans or objects
- Safety monitoring for text-to-motion generation models in physical robot control
- Preventing unsafe interactions in multi-agent or human-robot interaction scenarios
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.