Agile whole-body robot control learns safe movements without online filters
LIMBO: Learning and Internalizing Model-Free Barrier Objectives for Agile and Safe Whole-Body Control
Robotics
Summary
Controlling a robot's whole body safely is hard because it must avoid collisions and keep balance at the same time. The authors developed LIMBO, a system that learns to recognize safe and unsafe actions by watching how the robot reacts to different moves. This lets the robot create safer movement plans without needing extra safety checks while working. They tested LIMBO on a complex humanoid robot dodging balls and moving under obstacles and found it can handle those safely and smoothly.
What this means in practice
- •For robotics engineers: Create safer and more agile whole-body control policies for humanoid robots operating in cluttered or unpredictable environments.
- •For industrial automation teams: Implement learned safety criteria in robot control systems to reduce the need for external safety filters in automated manufacturing tasks.
Authors
Jake Gonzales, Arturo Flores Alvarez, Yu-Ming Chen, Aaron D. Ames, Lillian J. Ratliff, Manikantan Nambi
Abstract
Safe whole-body control requires coordinating collision avoidance and balance under high-dimensional, nonlinear dynamics--making safety certificates difficult to design and reuse across behaviors. We present LIMBO, a framework for synthesizing a state-action control barrier function and distilling its safety structure into a task policy. LIMBO learns the safety certificate from black-box transitions and a state-based failure specification over residual actions around a frozen base controller, making Q-CBF synthesis tractable in the full control dimension while placing the certificate in the task policy's control space. During synthesis, the learned safety value drives risk-guided sampling near the estimated boundary of recoverability; during task learning, it serves as a teacher that provides action-level safety feedback, yielding a robust task policy and alleviating the need for an online safety filter at deployment. We demonstrate LIMBO on a 29-degree-of-freedom humanoid performing dodgeball avoidance and locomotion beneath low obstacles. Beyond scaling learned Q-CBFs to whole-body control, we show that risk-guided boundary sampling provides a theoretically grounded way to explore the edge of recoverability. Under the same safety specification, ceteris paribus, varying the sampling concentration produces strategies ranging from crouching to a novel backward-leaning limbo maneuver. In both settings, the learned policies transfer to hardware without online safety filtering, showing that learned safety synthesis scales to agile whole-body control.