Safe, Task-Consistent Manipulation with Operational Space Control Barrier Functions

Daniel Morton and Marco Pavone
Stanford University
Paper arXiv Video Code CBFpy

Abstract

Safe real-time control of robotic manipulators in unstructured environments requires handling numerous safety constraints without compromising task performance. Traditional approaches, such as artificial potential fields (APFs), suffer from local minima, oscillations, and limited scalability, while model predictive control (MPC) can be computationally expensive. Control barrier functions (CBFs) offer a promising alternative due to their high level of robustness and low computational cost, but these safety filters must be carefully designed to avoid significant reductions in the overall performance of the manipulator. In this work, we introduce an Operational Space Control Barrier Function (OSCBF) framework that integrates safety constraints while preserving task-consistent behavior. Our approach scales to hundreds of simultaneous constraints while retaining real-time control rates, ensuring collision avoidance, singularity prevention, and workspace containment even in highly cluttered and dynamic settings. By explicitly accounting for the task hierarchy in the CBF objective, we prevent degraded performance across both joint-space and operational-space tasks, when at the limit of safety.

Overview

When enforcing safety, many trajectories can satisfy constraints

"Keep the tip of the end-effector in the safe-set"

❌ Bad operational space tracking

❌ Bad null space tracking

✅ Task-consistent

But not all are task-consistent

(minimally impeding the robot's performance while remaining safe)

Introducing

Operational Space Control Barrier Functions

Optimally balancing safety and performance for manipulators...
  • Across hundreds of safety constraints
  • At over kilohertz control rates
  • Without requiring workstation hardware
  • With open-source software in Python

We demonstrate a variety of common safety constraints

All of which can be enforced simultaneously at over kilohertz control rates

Constraint evolution for all safety conditions

For all constraints, we can reliably reach (and not exceed) the boundary of safety without over-conservative behavior.

Shown above: one teleoperated trajectory, with simultaneous constraint enforcement. Across 168 constraints, the robot remains safe (h > 0) without compromising on performance near the boundary of safety.

We then scale up to highly-cluttered scenes

And maintain kilohertz control rates even with over 400 collision-checking constraints.

By evaluating the full dynamics of the manipulator, we also maintain high performance even under input constraints and dynamic motion.

❌ Input-constrained kinematic model: Cannot account for torque limits

✅ Input-constrained dynamic model: Retains good tracking even under torque limits

Software

Our high-performance and easy-to-use Python package for CBFs is available on PyPI along with full documentation 

pip install cbfpy

and the OSCBF implementation is available at this project's GitHub repository.

BibTeX

@article{morton2025oscbf,
      author = {Morton, Daniel and Pavone, Marco},
      title = {Safe, Task-Consistent Manipulation with Operational Space Control Barrier Functions},
      year = {2025},
      journal = {arXiv preprint arXiv:2503.06736},
      note = {Submitted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Hangzhou, 2025},
      }