Straightness measurement of conveyors based on SINS/UWB with a Robust Laplace Kalman filter

Abstract

This study presents a robust adaptive Laplace Kalman filter (Robust-Laplace-KF) for enhancing the straightness measurement of scraper conveyors using SINS/UWB technology. By integrating the Laplace distribution into both process and measurement noise modeling, the Robust-Laplace-KF significantly improves robustness in heavy-tailed noise environments. The dual estimation of noise parameter scaling matrices and the one-step predictive probability density’s location vector facilitates comprehensive adaptive refinement of system state predictions. Extensive simulations and experimental analyses demonstrate the Robust-Laplace-KF’s superior performance under extreme conditions, including non-Gaussian noise and intense vibrations. The proposed method demonstrates higher accuracy and stability compared to existing techniques, with simulation results showing a reduction in measurement error from 55.482 cm to 3.613 cm under non-Gaussian noise conditions. Experimental validation using a scaled-down model highlights the method’s practical applicability, achieving a mean absolute error of 3.558 cm in straight configurations and 5.379 cm in oblique cutting configurations, representing a significant reduction in measurement errors. This research advances conveyor monitoring technologies, offering a promising solution for improving operational efficiency, reliability, and safety in mining and industrial applications.