Event
FTNCBF: A Fault-Tolerant Control Method Under Sensor Fault Attacks
1 reports · 1 sources
To address the failure of safety guarantees of neural control barrier functions when sensors are attacked, researchers propose the fault-tolerant neural control barrier function FTNCBF: estimators are run in parallel for each fault hypothesis, the estimation uncertainty is converted into first-order and second-order correction terms of the barrier condition, and fault-tolerant control is then synthesized via a minimum-norm program. The method has been reproduced in spacecraft and unmanned aerial vehicle scenarios, but the report does not provide quantitative results or comparisons for the safety guarantees.
Report timeline
- FTNCBF: Fault-Tolerant Neural Control Barrier Functions Under Sensor Failures and Attacks
FTNCBF addresses the failure of barrier-function safety guarantees when sensors are attacked. It runs estimators in parallel for each fault hypothesis, converts estimation uncertainty into first- and second-order correction terms in the barrier condition, and then synthesizes fault-tolerant control via a minimum-norm program. The method was reproduced in spacecraft and UAV scenarios.
Guyuehome ·