In art theory, negative volume—often referred to as negative space or voids—is the empty or open space around and between the subjects of an artwork. Instead of focusing on the subject itself (the "positive space" or mass), negative volume focuses on the shapes created by the surrounding area, making the "air" around an object just as crucial to the composition as the object itself. This is one way to build the world model and teach robots understand how to make a plan about their path.

In visual arts, the concept of negative volume (also known as negative space or voids in figure-ground theory) refers to the empty, unoccupied space surrounding and between the main subjects. Applying this to robotics: Negative volume represents the traversable medium — the open space that the robot’s body can actually move through, rather than focusing on cataloging endless obstacles. My unified framework leverages a single physical variable — the angle θ between surface normal and gravity vector (plus friction coefficient) — to simultaneously determine: Foothold / contact feasibility, and Negative volume (body passage constraints). This medium-centric, physics-grounded approach reduces semantic noise and offers a cleaner prior for building predictive world models in embodied AI and robotics. This art-inspired shift from ‘obstacle detection’ to ‘negative volume assessment’ provides a more holistic and generalizable representation for robot perception.

Wait, what am I talking about? Check the demo video right here! x.com/TentapC/status/2048116…

It’s finally here! 🎬 I spent some time perfecting the "Subtraction" logic. No more fighting with gravity. Check out the full solution below! 👇

See how they work tentapc.ca/products