A Novel Vehicle Localization Method Based-on Adaptive Singular Spectrum Analysis Using Low-Cost INS/GNSS

A low-cost inertial navigation system (INS) and global satellite (GNSS) fusion position estimator is affected by accuracy limitations multiple noises, leading to significant errors in positioning estimation. This paper proposes a new algorithm, SSA-ESKF, which combines singular spectrum analysis (SSA) an error-state Kalman filter (ESKF). The measurement unit (IMU) GNSS data obtained from the receivers are separately subjected SSA noise reduction. noise-reduced then utilized ESKF. Consequently, SSA-ESKF demonstrates superior performance terms of lower state compared conventional approach helps minimize impact caused neglecting higher-order Taylor expansion enhances linearization ESKF, thereby achieving improved accuracy. However, typically relies on empirically selecting constant values, may result incomplete or excessive separation signal noise. To address this limitation, we further propose adaptive spectral singularity (ASSA) that yields better results when integrating into verify proposed method, KITTI dataset experiments real vehicle with INS/GNSS were designed. comparison experimental between KF, ESKF ASSA-ESKF indicates superiority ASSA-ESKF. In addition, ablation conducted effectiveness IMU GPS independently, showed data.