Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a hierarchical control framework for humanoid robots that uses input-to-state safe control barrier functions (ISSf-CBFs) to ensure safety in the presence of disturbances. The framework integrates a kinematic-level whole-body controller, an ISSf-CBF safety filter, and a dynamic-level whole-body controller to prioritize tasks while guaranteeing kinematic safety under bounded disturbances. Real-robot experiments demonstrate improved safety margins and reliable enforcement of multiple safety constraints during locomotion, teleoperation, and balancing.
Humanoid robots can now maintain safety guarantees even with disturbances, thanks to a novel control barrier function approach that bridges kinematic and dynamic control.
Safety-critical control is essential for humanoid robots operating in complex human-centered environments, where physical safety constraints such as joint limits, self-collision avoidance, obstacle avoidance, and workspace boundaries must be satisfied during real-robot operation. However, existing approaches remain limited because kinematic safety guarantees can be degraded in the presence of unknown disturbances, such as model uncertainties, trajectory-tracking errors, and external perturbations. This paper presents a hierarchical safety-critical whole-body control framework for humanoid robots based on input-to-state safe control barrier functions (ISSf-CBFs). The proposed architecture integrates a kinematic-level whole-body controller (KinWBC), an ISSf-CBF safety filter, and a dynamic-level whole-body controller (DynWBC). KinWBC generates nominal joint-motion references from prioritized tasks; the ISSf-CBF filter minimally modifies these references to satisfy kinematic safety constraints under bounded disturbances; and DynWBC tracks the filtered references while enforcing full-body dynamic feasibility and contact stability. Safety constraints are imposed on a whole-body kinematic model, and the ISSf-CBF parameters are conservatively tuned so that the resulting kinematic safety guarantees can be transferred to full-order humanoid dynamics under unknown disturbances. Simulation and real-robot experiments demonstrate that the proposed framework improves safety margins under model mismatch and reliably enforces multiple safety constraints in real time during locomotion, teleoperation, and single-leg balancing with hand control. Project website: https://kwlee365.github.io/SafeWBC-Website/