aiqbench analysis
The paper learns the dynamics, nominal controller, and barrier function of an integral control barrier function filter from data, using neural ODEs, imitation learning, and supervised regression. It reports zero unsafe episodes across 200 held-out racing layouts at 100+ Hz on one CPU core, and a 53.38% constraint-satisfaction rate for a jointly trained baseline. These are results on one simulated 6-D bicycle benchmark, so they are not yet evidence for broader deployment, but the viability-based barrier labels and bounded input drift are useful ideas for anyone combining learned models with formal safety guarantees.
Neural Integral Control Barrier Functions for Agile Collision Avoidance
arXiv:2611.02626v1 Announce Type: cross Abstract: Control barrier functions (CBFs) are a standard way to keep a planner inside a safe set, but they assume control affine dynamics, a barrier function given in closed form, and a known nominal controller. Integral control barrier functions (ICBFs) remove the first assumption and absorb input constraints, yet they still assume a given nominal controller and a barrier function. In this work, we present an approach to learn all three from data, using neural ordinary differential equations (ODEs) for the dynamics, imitation for the nominal controller, and supervised regression for the barrier function. The learned dynamics, nominal controller, and the barrier function are used to build the ICBF filter constraint analytically. Furthermore, the barrier is trained on viability rather than geometry, meaning that a state counts as safe only if some emergency maneuver can still avoid the constraint under the actuator limits, a correction that geometric labels miss entirely and that grows with speed. Actuator limits then enter the online program as hard bounds while the learned barrier rows may be relaxed, which keeps the program solvable at every state and the actuator limits exact, and we bound how far the input it returns can drift from the one the true models would give. On agile racing with a 6-D dynamic bicycle, the filter removes all 12 unsafe episodes from 200 held out layouts at an unchanged success rate and above 100 Hz on one CPU core, holds the input constraint on every episode of a benchmark where a jointly trained baseline holds it on 53.38%, and keeps that advantage under large errors in the plant friction.
Abstract:Control barrier functions (CBFs) are a standard way to keep a planner inside a safe set, but they assume control affine dynamics, a barrier function given in closed form, and a known nominal controller. Integral control barrier functions (ICBFs) remove the first assumption and absorb input constraints, yet they still assume a given nominal controller and a barrier function. In this work, we present an approach to learn all three from data, using neural ordinary differential equations (ODEs) for the dynamics, imitation for the nominal controller, and supervised regression for the barrier function. The learned dynamics, nominal controller, and the barrier function are used to build the ICBF filter constraint analytically. Furthermore, the barrier is trained on viability rather than geometry, meaning that a state counts as safe only if some emergency maneuver can still avoid the constraint under the actuator limits, a correction that geometric labels miss entirely and that grows with speed. Actuator limits then enter the online program as hard bounds while the learned barrier rows may be relaxed, which keeps the program solvable at every state and the actuator limits exact, and we bound how far the input it returns can drift from the one the true models would give. On agile racing with a 6-D dynamic bicycle, the filter removes all 12 unsafe episodes from 200 held out layouts at an unchanged success rate and above 100 Hz on one CPU core, holds the input constraint on every episode of a benchmark where a jointly trained baseline holds it on 53.38%, and keeps that advantage under large errors in the plant friction.
Robotics (cs.RO); Machine Learning (cs.LG); Systems and Control (eess.SY)
arXiv:2611.02626 [cs.RO]
(or arXiv:2611.02626v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2611.02626 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 367 words
Related coverage
ReBound: Reset-Free Reinforcement Learning for Agile Driving via Reset-Aware Semi-Markov Bootstrapping
arXiv · Oct 7, 2026
Tight Time-Space Lower Bounds for Collision Finding and Element Distinctness under Label Symmetry
arXiv · Sep 11, 2026
A Modular Approach to Succinct Arguments for QMA
arXiv · Jun 10, 2026
Vision-based Underwater Formation Control With Input Saturations via Barrier Lyapunov Functions
arXiv · Sep 24, 2026