A Model-Based Spiking Neural Network Design for Real-time Spectra Analysis Using LMS Algorithm

• Published in: International Conference on Artificial Intelligence and Big Data (Online) pp. 678 - 683
• Main Author: Zhang, Lei
• Format: Conference Proceeding
• Language: English
• Published: IEEE 23.05.2025
• Subjects: Analytical models; Discrete Fourier Transform (DFT; Discrete Fourier transforms; Field programmable gate arrays; Least Mean Square (LMS); Mathematical models; Real-time systems; Signal processing; Signal processing algorithms; Spectral analysis; spectrum analysis; Spiking Neural Network (SNN); Spiking neural networks; Training
• ISSN: 2769-3554
• DOI: 10.1109/ICAIBD64986.2025.11082032

This paper presents a model-based phasor design for the estimation of Discrete Fourier Transform (DFT) coefficients utilizing the Least Mean Square (LMS) algorithm. Building upon the foundational connection between the LMS algorithm and the DFT, this study establishes a functional resemblance between the DFT and a spiking neural network (SNN) architecture. Furthermore, the paper explores the training of the SNN using the LMS algorithm. All simulations in this research were performed on a MATLAB-SIMULINK based platform. The proposed model integrates SNN architecture with DFT for spectrum analysis and employs the LMS learning technique to enhance the speed and efficiency of signal processing. This approach provides valuable insights for developing fast and efficient training solutions for artificial intelligence (AI).