Multi-Sensor Information Fusion Positioning of AUKF Maglev Trains Based on Self-Corrected Weighting

Sensors, Год журнала: 2025, Номер 25(8), С. 2628 - 2628

Опубликована: Апрель 21, 2025

Achieving accurate positioning of maglev trains is one the key technologies for safe operation and train schedules. Aiming at magnetic levitation positioning, there are problems such as being easily interfered with by external noise, single method, traditional weighting affected historical data, which lead to deviation fusion results. Therefore, this paper adopts self-corrected Sage–Husa noise estimation algorithms improve them proposes a research method multi-sensor information an AUKF based on self-correcting weighting. Multi-sensor technology applied trains, does not rely sensor. It combines algorithm unscented Kalman filter (UKF) form using data collected cross-sensor lines, INS, Doppler radar, GNSS, adaptively updates statistical feature measurement eliminates single-function low-integration shortcomings various modules achieve precise trains. The experimental results point out that self-correction-based trajectories closer real values, their ME RMSE errors smaller, indicating self-correction-weighted proposed in has significant advantages terms stability, accuracy, simplicity.

Язык: Английский

An Integrated Navigation Algorithm for Underwater Vehicles Improved by a Variational Bayesian and Minimum Mixed Error Entropy Unscented Kalman Filter

Electronics, Год журнала: 2024, Номер 13(23), С. 4727 - 4727

Опубликована: Ноя. 29, 2024

In complex marine environments, autonomous underwater vehicles (AUVs) rely on robust navigation and positioning. Traditional algorithms face challenges from sensor outliers non-Gaussian noise, leading to significant prediction errors filter divergence. Outliers in observations also impact positioning accuracy. The original unscented Kalman (UKF) based the minimum mean square error (MMSE) criterion suffers performance degradation under these conditions. This paper enhances entropy algorithm using variational Bayesian (VB) methods mixed functions. By implementing (MEE) kernel functions UKF, algorithm’s robustness conditions is improved. VB method adaptively fits measurement noise covariance, enhancing adaptability environments. Simulations sea trials validate proposed performance, showing improvements accuracy root (RMSE). environments with our improves overall by at least 10% over other existing algorithms. demonstrates high of algorithm.

Язык: Английский