Application and Performance Assessment of Velocity Correction Filter under Low-Velocity Regime

The case analysis of an extended Kalman filter (EKF) application presented in this paper improves the quality of velocity measurements using a horizontal acoustic Doppler current profiler (H-ADCP). Field measurements were conducted at three river sites: Gwangjin Bridge, Yeoju Bridge, and Hoguk Bridge. Rather than applying the filter across the entire dataset, EKF was selectively applied to segments with coefficient of variation (CV) exceeding a predefined threshold, which indicated unstable velocity patterns. In the absence of a reference velocity, the filtering performance was evaluated based on changes in CV rather than conventional metrics such as or root mean square error (RMSE). The results demonstrate that EKF effectively reduces the signal variability in unstable segments while preserving the physical characteristics of the time series. This research highlights the potential of selective EKF-based filtering in enhancing the reliability of in-situ velocity measurements and provides a basis for developing real-time quality assessment and automatic correction systems in hydraulic monitoring.