Neural network-based direct robust adaptive non-fragile fault-tolerant control of amorphous flattened air-ground wireless self-assembly system

Abstract

Purpose In this paper, the authors take an amorphous flattened air-ground wireless self-assembling network system as the research object and focus on solving the wireless self-assembling network topology instability problem caused by unknown control communication faults during the operation of this system. Design/methodology/approach In the paper, the authors propose a neural network-based direct robust adaptive non-fragile fault-tolerant control algorithm suitable for the air-ground integrated wireless ad hoc network integrated system. Findings The simulation results show that the system eventually tends to be asymptotically stable, and the estimation error asymptotically tends to zero with the feedback adjustment of the designed controller. The system as a whole has good fault tolerance performance and autonomous learning approximation performance. The experimental results show that the wireless self-assembled network topology has good stability performance and can change flexibly and adaptively with scene changes. The stability performance of the wireless self-assembled network topology is improved by 66.7% at maximum. Research limitations/implications The research results may lack generalisability because of the chosen research approach. Therefore, researchers are encouraged to test the proposed propositions further. Originality/value This paper designs a direct, robust, non-fragile adaptive neural network fault-tolerant controller based on the Lyapunov stability principle and neural network learning capability. By directly optimizing the feedback matrix K to approximate the robust fault-tolerant correction factor, the neural network adaptive adjustment factor enables the system as a whole to resist unknown control and communication failures during operation, thus achieving the goal of stable wireless self-assembled network topology.