A New Zeroing Neural Network with Tolerance to Harmonic Noise for Solving Time-Variant Linear Equation

Recently, a typical recurrent neural network termed the zeroing neural network (ZNN) with prominent properties (e.g., finite-time convergence and noise suppression) has been studied to solve the time-variant linear equation. However, these ZNN models may not perform well when being affected by harmonic noise. In this paper, for completeness, a new ZNN model that is tolerant to harmonic noise is proposed and investigated for solving the time-variant linear equation. Such a model is then theoretically analyzed to converge the computational error to zero, thereby ensuring the model convergence property. Simulation results with three examples are illustrated to further validate the efficacy of the proposed harmonic-noise-tolerance ZNN model.