Adaptive Neural Prescribed Performance Control for Non-Triangular Structural Stochastic Highly Nonlinear Systems Under Hybrid Attacks

In this article, the adaptive neural network prescribed performance control issue is followed with interest for a class of non-triangular structural stochastic highly nonlinear interconnected systems under hybrid attacks, that is, injection attack and deception attack. Unlike the previous achievements, the original states of the system are unknown under the influence of hybrid attacks, and it is particularly difficult to handle highly nonlinear interconnection functions of non-triangular structure in the system. Hence, an effective variable separation method based on adaptive compensation technology is constructed, such that the complex nonlinear functions can be separated and further simplified, and then approximated by neural network. By utilizing a prescribed performance function, the tracking error of the system can converge in a fixed time based on the Lyapunov stochastic stability theory and the framework of the backstepping technology. Ultimately, the reliability and effectiveness of the proposed control strategy can be verified by a practical example.