Dual-Channel NCSs Performance Error Estimation Under DoS Attacks and Intelligent Control Supervised by Machine Learning to AGV Application
Xiao Cai, Kaibo Shi, Yanbin Sun, Jinde Cao, Shiping Wen, Chen Peng, Zhihong Tian
Abstract
This study focuses on addressing the issue of estimating performance error (PE) in dual-channel networked control systems (NCSs) under DoS attacks. Firstly, the study examines the impact of network congestion caused by quality of service (QoS) management mechanisms and denial-of-service (DoS) attacks. A signal compression mechanism is proposed to mitigate network congestion. Additionally, an asymmetric Lyapunov-Krasovskii function (LKF) is introduced to reduce decision variables, and it is split into two parts to compensate for the asymmetry and increase the LKF energy, which approach helps to reduce the conservatism of the criterion. Furthermore, an integral-based event-triggered mechanism (IETM) supervised by a machine learning algorithm is developed to estimate the PE of NCSs. Finally, the effectiveness of the proposed method is demonstrated through verification on the automated ground vehicle (AGV) CarSim-Simulink joint platform, confirming its feasibility.