A Robust Cubature Kalman Filter with Abnormal Observations Identification Using the Mahalanobis Distance Criterion...

Area of Science:

• Robotics and Control Systems
• Navigation and Positioning Technologies
• Signal Processing

Background:

• Inertial Navigation System/Global Navigation Satellite System (INS/GNSS) integration is crucial for intelligent transportation systems.
• GNSS signal blockage in urban/suburban areas degrades INS/GNSS navigation performance due to observation uncertainty.
• Existing methods struggle with the nonlinearities and uncertainties inherent in INS/GNSS integration.

Purpose of the Study:

• To develop a novel robust Cubature Kalman Filter (CKF) for improved INS/GNSS integration in vehicular navigation.
• To address the challenge of abnormal GNSS observations caused by signal blockage.
• To enhance the robustness and reliability of navigation solutions in complex environments.

Main Methods:

• A robust CKF incorporating a scaling factor derived from the Mahalanobis distance criterion was developed.
• A theory for identifying abnormal observations using the Mahalanobis distance criterion was established.
• The Mahalanobis distance criterion was used to calculate a scaling factor that inflates observation noise covariance, reducing filtering gain during abnormal observations.

Main Results:

• The proposed robust CKF effectively identifies and mitigates the influence of abnormal GNSS observations.
• The filter demonstrated improved robustness in INS/GNSS integration, particularly in scenarios with signal blockage.
• Simulation and experimental results validated the enhanced performance of the robust CKF for vehicular navigation.

Conclusions:

• The novel robust CKF significantly improves the reliability of INS/GNSS integration for vehicular navigation.
• The Mahalanobis distance criterion provides an effective mechanism for handling abnormal observations.
• This approach offers a promising solution for robust navigation in intelligent transportation systems.