Geomagnetic/Inertial Fusion Indoor Navigation System Based on an Improved WD-UKF*

Accurate indoor positioning has become increasingly important as the demand for high-precision navigation grows. However, traditional single-sensor technologies struggle to meet these requirements. This paper presents an enhanced geomagnetic/inertial fusion navigation system based on a Wavelet Denoising-Unscented Kalman Filter (WD-UKF). The system integrates geomagnetic and inertial data, decomposing the signals into multi-frequency components using wavelet transform, followed by denoising through an optimized thresholding function. Nonlinear state estimation of the denoised components is performed using the Unscented Kalman Filter (UKF). Evaluations in complex indoor environments show that the proposed WD-UKF significantly mitigates noise and improves positioning accuracy, achieving a maximum spindle turning error of only 3.7 cm, These results demonstrate the system's robustness and precision, making it well-suited for dynamic and challenging indoor navigation scenarios.