Uncertainty-Aware Safety-Critical Control: Design and Application
针对工业信息物理系统中模型误差、工况变化等不确定性带来的安全挑战，提出一种基于屏障函数的安全关键控制方法，利用高斯过程建模不确定性并融入控制屏障函数，提升系统鲁棒性。
Cyber-physical systems (CPSs) in industrial environments are expected to achieve high control performance while satisfying strict safety constraints. However, practical industrial systems often involve significant uncertainties arising from modeling errors, varying operating conditions, and external disturbances, which pose significant challenges to traditional control methods in achieving efficient and safe operation. To address these challenges, this article proposes a barrier function-based safety-critical control (BF-SCC) method. Specifically, to compensate for the limited capability of nominal models in representing system dynamics under uncertainties, a Gaussian process (GP)-based uncertainty modeling method is first proposed to capture the modeling mismatch between the nominal model and the actual system. Then, by incorporating uncertainty information into the control barrier function (CBF), an uncertainty-aware barrier function method is developed to improve system robustness. Moreover, to adapt to time-varying uncertainties in multiple operating conditions, an event-triggered update mechanism for the GP model is developed to continuously refine the uncertainty representation. Compared with existing safety-critical control methods, the proposed BF-SCC method better handles uncertainties while ensuring efficient system operation under safety constraints.