Adaptive neural output-feedback safe control for high-relative-degree uncertain nonlinear systems

This paper proposes an adaptive neural output-feedback control scheme that guarantees both safety and tracking for high-relative-degree uncertain nonlinear systems. A high-gain observer (HGO) and two adaptive radial basis function neural networks (RBFNNs) are developed for joint state estimation and dynamics learning. The proposed estimation–learning mechanism enables the use of lower observer gains, thereby mitigating the peaking phenomenon. A baseline controller is then constructed based on filtered tracking error dynamics to ensure tracking performance, while output safety is enforced through an exponential control barrier function-based quadratic program that minimally modifies the baseline control input. Comparative simulation results demonstrate the effectiveness of the proposed method.