Noise Reduction Techniques for Sensor Data: Comparative Analysis of Kalman, Butterworth, Savitzky-Golay, Median, and Moving Average Filters for UWB-Based Position Estimation

Authors : Levent Türkler, Lütfiye Özlem Akkan

Pages : 146-159

Doi:10.18466/cbayarfbe.1682594

View : 41 | Download : 99

Publication Date : 2025-12-29

Article Type : Research Paper

Abstract :Accurate localization of autonomous mobile systems has become a critical requirement in modern engineering applications. However, field environments often lead to erroneous position data due to signal interference or unexpected behaviors of signals in the presence of obstacles. In this study, raw data obtained from an Ultra-Wideband (UWB) positioning system was intentionally degraded by amplification and the addition of artificial noise to simulate realistic signal corruption. Subsequently, five different filters were evaluated to denoise this highly contaminated data. The performance of each filter was tested using custom-developed software by comparing its unoptimized and optimized configurations. As a key outcome, the optimal parameter set of the most effective filter for noise reduction was identified and reported.
Keywords : Kalman, Butterworth, Savitzky-Golay, Median, Filters, MAF

ORIGINAL ARTICLE URL