Adaptive stabilization of strict-feedback nonlinear systems using the fully actuated system approach and neural network approximation
This paper studies the adaptive stabilization of a class of strict-feedback systems (SFSs) with unknown nonlinearities and bounded matched disturbances. Using the fully actuated system (FAS) approach, the strict-feedback system is transformed into a high-order form with a single lumped unknown nonlinear term and a bounded disturbance, which leads to a more compact controller design. An Radial basis function (RBF) neural network (NN) is used to approximate the lumped unknown nonlinearity online, while a smooth robust term is introduced to compensate for the approximation error and the external disturbance. An adaptive controller is then developed based on this formulation. Lyapunov analysis shows that the transformed state and the neural network weight-estimation error are uniformly ultimately bounded. The original system states ultimately enter and remain within a neighborhood of the origin, while the neural network weight estimate and the control input remain bounded. The proposed method is further compared with an adaptive RBF neural network backstepping controller in a numerical example and an resistor-inductor-capacitor (RLC) circuit. In both cases, the proposed method achieves lower root-mean-square error, steady-state root-mean-square error, and control effort.
Authors
- Da‐Ke Gu (ORCID: https://orcid.org/0000-0003-0611-5004)
- Yindong Liu (ORCID: https://orcid.org/0000-0003-0212-7980)
- Yuxin Deng (ORCID: https://orcid.org/0009-0003-0581-3633)
Institutions
- Northeast Electric Power University (CN)
Publication Details
- Journal
- Measurement and Control
- Published
- 2026-11-21
- DOI
- https://doi.org/10.1177/00202940261511863
- Primary Topic
- Adaptive Control of Nonlinear Systems
- Type
- article
- Field-Weighted Citation Impact
- 0.00