Clock-Shift-Aided Blind Doppler Matching in Passive SDR Monitoring Data for LEO Doppler Positioning

Abstract

Low Earth orbit (LEO) satellite signals have recently attracted increasing attention as opportunistic signals for positioning, navigation, and timing. Most Doppler-based LEO positioning studies rely on tracked satellite signals or raw in-phase/quadrature (IQ) samples that preserve signal-level structures for acquisition, Doppler tracking, or satellite association. In contrast, passive software-defined radio (SDR) monitoring systems often retain only time-frequency power scans, in which LEO Doppler traces appear as unlabeled ridges without raw IQ samples or identifiable signal structures. This paper proposes a clock-shift-aided segment-based blind Doppler matching method that associates unlabeled Doppler traces with candidate satellites predicted using two-line element (TLE) data. The proposed framework consists of slope-consistent Doppler segment generation, clock-shift-aided temporal alignment, and multi-metric satellite matching based on Doppler-rate and trajectory-shape consistency. The proposed method was evaluated using a three-hour subset of open Starlink/OneWeb monitoring data collected in the United Kingdom. The clock-offset search selected an effective offset of +19 s, for which 56 of 69 segments had at least one slope-compatible candidate, compared with 19 segments at zero shift. Using the resulting satellite associations, the batch positioning error over the full three-hour period was 104.2 m, whereas the RMSE-based and slope-based simple-segment baselines failed to converge near the reference receiver location.