Certified Neural Network Control Architectures: Methodological Advances in Stability, Robustness, and Cross-Domain Applications

Rui Liu, Jianhua Huang (), Biao Lu () and Weili Ding
Additional contact information
Rui Liu: Institute of Robotics and Automatic Information System, College of Artificial Intelligence, Nankai University, Tianjin 300353, China
Jianhua Huang: State Key Laboratory of Intelligent Manufacturing of Advanced Construction Machinery, Jiangsu XCMG National Key Laboratory Technology Co., Ltd., Xuzhou 221004, China
Biao Lu: Institute of Robotics and Automatic Information System, College of Artificial Intelligence, Nankai University, Tianjin 300353, China
Weili Ding: National Key Laboratory of Key Technologies for Lifting Machinery, Yanshan University, Qinhuangdao 066004, China

Mathematics, 2025, vol. 13, issue 10, 1-45

Abstract: Neural network (NN)-based controllers have emerged as a paradigm-shifting approach in modern control systems, demonstrating unparalleled capabilities in governing nonlinear dynamical systems with inherent uncertainties. This comprehensive review systematically investigates the theoretical foundations and practical implementations of NN controllers through the prism of Lyapunov stability theory, NN controller frameworks, and robustness analysis. The review establishes that recurrent neural architectures inherently address time-delayed state compensation and disturbance rejection, achieving superior trajectory tracking performance compared to classical control strategies. By integrating imitation learning with barrier certificate constraints, the proposed methodology ensures provable closed-loop stability while maintaining safety-critical operation bounds. Experimental evaluations using chaotic system benchmarks confirm the exceptional modeling capacity of NN controllers in capturing complex dynamical behaviors, complemented by formal verification advances through reachability analysis techniques. Practical demonstrations in aerial robotics and intelligent transportation systems highlight the efficacy of controllers in real-world scenarios involving environmental uncertainties and multi-agent interactions. The theoretical framework synergizes data-driven learning with nonlinear control principles, introducing hybrid automata formulations for transient response analysis and adjoint sensitivity methods for network optimization. These innovations position NN controllers as a transformative technology in control engineering, offering fundamental advances in stability-guaranteed learning and topology optimization. Future research directions will emphasize the integration of physics-informed neural operators for distributed control systems and event-triggered implementations for resource-constrained applications, paving the way for next-generation intelligent control architectures.

Keywords: neural network controller; stability certification; robustness; formal verification; cross-domain applications (search for similar items in EconPapers)
JEL-codes: C (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2227-7390/13/10/1677/pdf (application/pdf)
https://www.mdpi.com/2227-7390/13/10/1677/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jmathe:v:13:y:2025:i:10:p:1677-:d:1660131

Access Statistics for this article

Mathematics is currently edited by Ms. Emma He

More articles in Mathematics from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().