Analyzing the Multipath Characteristics in GNSS Code Tracking Using Dynamic Mode Decomposition

Multipath interference from signal reflections remains a significant challenge for accurate Global Navigation Satellite System (GNSS) positioning in urban environments. This paper introduces Dynamic Mode Decomposition (DMD), a data-driven method for analyzing dynamic systems, to characterize the spatial and dynamic characteristics of multipath effects in GNSS code tracking. By treating a time series of GNSS autocorrelation function (ACF) outputs as a dynamic system, DMD decomposes the composite signal into distinct spatial modes and their associated temporal evolutions. In controlled simulation experiments involving a direct line-of-sight signal and a single time-varying reflected signal, DMD successfully isolates the two components. The results demonstrate that the extracted modes accurately correspond to the spatial shapes (amplitude and code delay) of the direct and reflected signals, while their eigenvalues correctly capture the direct signal's stability and the reflected signal's attenuation dynamics. This study establishes the potential of DMD as a novel tool for analyzing and decomposing GNSS multipath signals, providing a foundation for future advanced mitigation techniques in complex urban scenarios.