Comparative study of denoising techniques for 5G communication at 3.5 GHz

Abstract

This paper presents a comparative study of denoising techniques for improving 5G communication at 3.5GHz. A 5G system is simulated in MATLAB with thermal noise, intermodulation noise, and external interference. The wavelet, PCA, Wiener, median, and Kalman filters are evaluated using SNR, MSE, and PSNR. Results show that the Kalman filter achieves the best performance, reducing MSE by 92.5% and improving SNR by 14 dB over other techniques. Under thermal noise, it attained an SNR of 20.47 dB, while for all noise sources combined, it maintained 19.43 dB. Wavelet and median filters performed better under thermal noise, whereas PCA and Wiener filters were more effective for combined noise. These findings provide a quantitative basis for selecting optimal denoising techniques, aiding efficient 5G communication system design.