Neural Network-Based Nonconservative Predefined-Time Backstepping Control for Uncertain Strict-Feedback Nonlinear Systems

To handle the tracking problem of uncertain strict-feedback nonlinear systems with matched and mismatched composite disturbances, this article studies a predefined-time backstepping controller by resorting to a Lyapunov-based predefined-time dynamic paradigm, a regulation function, and neural networ...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transaction on neural networks and learning systems 2024-11, Vol.35 (11), p.16562-16573
Main Authors: Lv, Jixing, Ju, Xiaozhe, Wang, Changhong
Format: Article
Language:English
Subjects:
Adding-absolute-value (ADV) technique
Aerospace electronics
Artificial neural networks
Backstepping
Convergence
neural network (NN)
Nonlinear systems
predefined-time backstepping controller
Regulation
regulation function
uncertain strict-feedback nonlinear system
Uncertainty
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:To handle the tracking problem of uncertain strict-feedback nonlinear systems with matched and mismatched composite disturbances, this article studies a predefined-time backstepping controller by resorting to a Lyapunov-based predefined-time dynamic paradigm, a regulation function, and neural networks (NNs). Moreover, an adding-absolute-value (ADV) technique is adopted in the design process to remove the control singularity. Theoretical analyses prove the boundedness of all closed-loop system signals and the predefined-time convergence of the tracking error into an arbitrarily small vicinity of the origin. The proposed controller exhibits four advantages: 1) the actual convergence time is precisely predefined by only one design parameter irrespective of the initial conditions, and the control energy is economized; 2) no unbounded terms are adopted for predefining the actual convergence time, thus avoiding numerical overflow problem under limited memory space and gaining strong noise-tolerant ability; 3) the peaking tracking error and control input magnitude can be effectively reduced by appropriately setting parameters of the regulation function; and 4) the controller is continuous and nonsingular everywhere. Finally, a practical example of a single-link manipulator is presented to validate the efficacy and superiority of our predefined-time controller.
ISSN:2162-237X
2162-2388
2162-2388
DOI:10.1109/TNNLS.2023.3296194