Distributionally Robust Safe Control of Robotic Manipulators in Dynamic Environments

Update (17 March 2026)

Description

We implement and compare two Control Barrier Function (CBF) approaches for a UR3 manipulator avoiding moving spherical obstacles, using vision-based measurements with added Gaussian noise:

• Nominal CBF (src/nominal_CBF_single_trial.py, src/nominal_CBF_multiple_trials.py)
• Distributionally Robust CBF (src/DR_CBF_single_trial.py, src/DR_CBF_multiple_trials.py)

Requirements

• Python: tested with Python 3.13.5
• PyBullet: tested with PyBullet 3.2.7
• NumPy: tested with Mumpy 2.1.3
• SciPy: tested with SciPy 1.15.3

Acknowledgments

Parts of this project page were adopted from the Nerfies page.