A High-Precision PDR Positioning System with Heading Constraint Model using a Low-cost IMU

The Pedestrian Dead Reckoning (PDR) positioning system based on Inertial Measurement Units (IMUs) has attracted significant attention due to its independence from external infrastructure. However, conventional step-model-based PDR methods exhibit declining positioning accuracy over extended distances, limiting their practical application. To address these limitations, we propose a high-precision PDR system employing a low-cost IMU, integrated with a heading constraint model to enhance performance. An Extended Kalman Filter (EKF) is applied for heading estimation, augmented by a directional constraint model to improve pedestrian navigation accuracy. The framework further incorporates comprehensive gait phase identification criteria and an adaptive stride length estimation strategy derived from IMU data. Multi-group long-distance pedestrian walking experiments validate the algorithm's performance, demonstrating superior positioning accuracy (average position error of 0.36 m; start-end position error of 0.91 m).