ENHANCED ACCURACY IN PENETRATING POSITIONING USING UWB TECHNOLOGY BASED ON RECEIVED SIGNAL STRENGTH AND MACHINE LEARNING

Tóm tắt

This paper proposes a novel method to enhance accuracy in ultra-wideband penetrating positioning systems by using the raw data elimination technique combined with a machine learning model applied to received signal strength data. The emergence of ultra-wideband technology has addressed many challenges related to radio frequency spectrum scarcity, offering high precision in distance measurement and positioning. However, it still faces significant challenges such as multipath propagation, scattering, and refraction which degrade system performance. To address these issues, various signal processing approaches have been utilized, including machine learning techniques. In the proposed approach, an optimized LightGBM-based machine learning model is employed, which significantly improves the accuracy of ultra-wideband penetrating positioning systems. Computational results indicate that the proposed method reduces the mean absolute error by 28.2% to 72% compared to existing methods. This represents an effective research direction that addresses complex challenges in the field of radio-based localization and enhances the performance of both penetrating and indoor positioning systems.