Robust Pedestrian Dead Reckoning Based on MEMS-IMU for Smartphones

Area of Science:

• * Navigation Systems
• * Mobile Sensing
• * Human-Computer Interaction

Background:

• * Pedestrian dead reckoning (PDR) is crucial for indoor navigation.
• * Existing strap-down inertial navigation system (SINS)-based PDR algorithms face challenges with accuracy and robustness.
• * Smartphone sensors (gyros, accelerometers, magnetometers) offer a viable platform for PDR.

Purpose of the Study:

• * To develop a more robust PDR algorithm utilizing smartphone SINS.
• * To enhance position and velocity estimation accuracy compared to existing methods.
• * To evaluate the algorithm's performance across various smartphone usage scenarios.

Main Methods:

• * Integration of gait models and motion constraints for pseudo-measurements (3D velocity, 2D position increment).
• * Utilization of gravity vector, magnetic field vector, and quasi-static attitude.
• * Comparative analysis against existing SINS-based PDR and gait-model-based PDR algorithms.

Main Results:

• * The proposed algorithm demonstrates improved position estimation compared to existing SINS-based PDR.
• * Superior performance observed in specific smartphone positions: handheld, calling, swaying, and pocket.
• * Enhanced accuracy particularly noted when the smartphone is in a swaying hand and step detection fails.

Conclusions:

• * The advanced PDR algorithm offers a more robust and accurate solution for smartphone-based navigation.
• * The method effectively handles variations in smartphone orientation and sensor performance.
• * This research contributes to more reliable pedestrian navigation systems using ubiquitous mobile devices.