Noise Cancellation Berbasis Ai Dengan Algoritma Lms Untuk Mengurangi Gangguan Noise Pada Komunikasi Radio Trunking

Abstract

  The effectiveness of trunking radio communication systems in critical operations is often degraded by intense environmental noise, which increases the risk of fatal information errors. This study aims to enhance voice signal quality by designing an Active Noise Cancellation (ANC) system based on simple Artificial Intelligence (AI) utilizing the Least Mean Squares (LMS) algorithm. Unlike conventional static filters, the proposed method offers high adaptability to dynamic acoustic disturbances in the field. The research methodology involves computational simulation using the ADALINE neural network architecture, where clean speech signals are mixed with various operational noise profiles at a standard 8 kHz sampling frequency. System performance is objectively evaluated based on Signal-to-Noise Ratio (SNR) improvement, convergence speed, and Steady-State Error stability. The results indicate that the proposed system successfully increased the average SNR by 12 dB to 15 dB and achieved rapid convergence when noise characteristics changed abruptly. Further spectral analysis confirmed that the algorithm effectively attenuated disturbance energy without distorting human vocal frequencies. With low computational complexity, this approach proves efficient and feasible for implementation on resource-constrained radio hardware. In conclusion, the adaptive LMS algorithm provides a significant solution for clarifying trunking radio communication in real-time.by applying the SAW method, which can rank the assessment results.