Drift Reduction in Pedestrian Navigation System by Exploiting Motion Constraints and Magnetic Field

Area of Science:

• * Inertial Navigation Systems
• * Sensor Fusion
• * Pedestrian Navigation

Background:

• * Foot-mounted MEMS inertial sensors in Pedestrian Navigation Systems (PNS) rely on zero-velocity updates (ZUPTs) to mitigate drift and estimate sensor errors.
• * ZUPTs alone are insufficient for error reduction, particularly for unobservable heading errors, leading to position drift.
• * Existing methods necessitate additional motion constraints and heading reduction information for improved accuracy.

Purpose of the Study:

• * To exploit additional pedestrian gait motion constraints: walking along straight paths and standing still.
• * To develop a Magnetic Anomaly Detection (MAD) and compensation algorithm for robust heading estimation.
• * To reduce drift in zero-velocity updated Inertial Navigation Systems (INS) within GPS-denied and magnetically distorted environments.

Main Methods:

• * Integration of two novel motion constraints ('virtual sensors') representing straight path walking and prolonged standing.
• * Design and implementation of a Magnetic Anomaly Detection (MAD) and compensation algorithm.
• * Incorporation of only healthy magnetometer data into the Extended Kalman Filter (EKF) updating step.

Main Results:

• * The proposed motion constraints significantly reduce drift in pedestrian navigation solutions.
• * The MAD algorithm effectively filters out magnetic distortions, improving heading accuracy.
• * Experimental validation in GPS-denied and magnetically distorted environments demonstrates reduced INS drift.

Conclusions:

• * Combining gait motion constraints with a robust magnetic anomaly detection algorithm enhances PNS accuracy.
• * The developed approach offers a viable solution for reliable pedestrian navigation in challenging environments.
• * This work contributes to more precise and dependable indoor and GPS-denied navigation solutions.