Riding the Shifting Potential: When Reactive Control Suffices for Multi-Goal Behavior
2026-05-26 • Robotics
Robotics
AI summaryⓘ
The authors address problems where a robot has to handle multiple goals that sometimes clash, which normally causes simple reactive techniques to get stuck. They improve this by using a graph-based model of the environment and applying a math trick called nullspace projection to prioritize and resolve these conflicting goals dynamically. They tested this approach on tasks like guiding robots around tricky obstacles and pushing oddly shaped objects, achieving much better success than other methods without extra training data. Their method also worked well on a real robot, showing it can handle real-world constraints using the same idea.
reactive controlmulti-objective optimizationlocal minimanullspace projectiongraph-based world modelpotential fieldsnon-convex obstaclesrobot navigationplanar pushingprioritization
Authors
Vito Mengers, Oliver Brock
Abstract
Reactive control is often considered insufficient for multi-objective tasks because conflicting objectives give rise to local minima. We argue this limitation is not inherent but arises from static encodings that fail to reflect how objectives currently interact. We exploit the interaction structure encoded in a graph-based world model by extending it with nullspace projections: conflicts are resolved where they arise by projecting lower-priority gradients into the nullspace of higher-priority ones, with priorities determined continuously from the current state. We demonstrate this in two domains where conflicts between objectives are central: navigation around non-convex obstacles, where static potential fields fundamentally fail, and planar pushing of non-convex objects, where our method achieves $100\%$ success across one-hundred configurations versus $0\%$ for the steepest-descent baseline and ${\sim}55\%$ for diffusion policy, without demonstrations or retraining. The same formulation transfers directly to a real robot with additional perceptual and kinematic constraints, accommodating them through the same mechanism.