Safety Control of Uncertain Nonlinear Systems Based on a Neural Network-Based Control Barrier Function

The primary focus of this paper is on the safety control problem of uncertain nonlinear systems. Based on the idea of proving the invariance of the superlevel sets of scalar constraint functions with Control Barrier Functions (CBFs), we propose a novel neural network-based control barrier function (NNCBF) control scheme to achieve safety control of uncertain nonlinear systems with state constraints. According to the CBF and neural network (NN) technology, the uncertain system is reconstructed and the relative definitions of NNCBF are presented. Subsequently, by resolving a quadratic program (QP) using the NNCBF condition as a constraint, the nominal controller is adapted to achieve safety control without violating constraints. Finally, an illustrative example is performed to validate the efficacy of the proposed safety control method utilizing NNCBF - QP.