This item is under embargo and not available online per the author's request. For access information, please visit http://digital.wustl.edu/publish/etd-search.html.
Abstract
We address localization drift and sensor faults in humanoid robots controlled by safety-critical CLF–CBF-based controllers. The Unitree G1’s internal odometry exhibits significant drift during walking, which degrades obstacle-avoidance performance and can invalidate formal safety guarantees when combined with faulty sensors. To obtain drift-free state estimates, we design an Extended Kalman Filter (EKF) that fuses IMU and joint encoder measurements with the built-in odometer in a discrete-time single-integrator model of the robot’s center-of-mass motion. On top of this estimator, we develop a nested Control Barrier Function (CBF)–based fault-tolerant control architecture. A bank of observers generates residuals for multiple fault hypotheses (nominal and biased sensors). For each hypothesis we construct a CBF defining a corresponding safe set and use observer error bounds to select an active fault mode. The control input is then synthesized from a quadratic program enforcing the CBF constraint for the selected mode, thereby tightening safety margins when more severe faults are suspected. Simulation studies with position bias faults on different state channels show that the EKF significantly reduces localization error and that the nested-CBF fault-tolerant controller maintains collision avoidance and set invariance even in the presence of persistent sensor biases.
Document Type
Article
Class Name
Electrical and Systems Engineering Undergraduate Research
Language
English (en)
Date of Submission
12-8-2025
Recommended Citation
Peng, Ziyi; Cox, Jackson; and Clark, Andrew, "Humanoid Localization and Fault Tolerant Control via Nested Control Barrier Functions" (2025). Electrical and Systems Engineering Undergraduate and Graduate Research. 51.
https://openscholarship.wustl.edu/eseundergraduate_research/51