Prescribed performance control for nonlinear multiagent systems with information protection mechanism

Abstract

This paper is dedicated to the prescribed performance control issue for nonlinear multiagent systems with information protection mechanism. When the system encounters complex environmental changes, if the constraint boundaries cannot be flexibly adjusted according to the external environment, it will lead to a decrease in system performance. To address this issue, we propose a novel time-dependent shift function, which is combined with the traditional barrier Lyapunov function method. This combination allows the tracking errors to converge to different boundaries in different time intervals, thereby improving the robustness of the system in dynamic environments. At the same time, a custom information protection mechanism is designed to enhance the security of the system. This mechanism uses an output mask function with adjustable time intervals to protect information freely determined by the user. In addition, it can maintain system performance without decryption, avoiding the computational burden and information leakage issues caused by the decryption process. To handle the unknown uncertain terms in the controlled systems, neural network (NN) is used for identification. An improved cooperative learning control protocol is proposed to enhance the generalization ability of NN, which removes the assumption on neighbor adaptive law in existing cooperative learning literature. All signals of the closed-loop system are bounded. Finally, two simulation examples verify the effectiveness of the presented approach.