Research on Pedestrian Navigation Algorithm Based on Zero Velocity Update/Heading Error Self-observation/Geomagnetic Matching

Nowadays, pedestrian navigation technology is playing an increasingly important role in supermarket shopping, fire rescue and field exploration, and the pedestrian navigation and positioning without global navigation satellite system (GNSS) has become an indispensable link. The self-contained sensors are used as a hardware platform for research on pedestrian autonomous navigation in non-GNSS environment. A zero-speed comprehensive discriminant algorithm based on the 2+2 hierarchical model is studied to improve the accuracy and reliability of zero velocity update (ZUPT). Kalman filter algorithm based on ZUPT designed for inertial navigation system is used to effectively suppress the sensor error divergence. To solve the problem of pedestrian long-term heading divergence, the magnetic heading error self-observation algorithm (MHESO) for pedestrian initial static state and the ZUPT heading error self-observation algorithm (ZHESO) for pedestrian movement are studied. In addition, a geomagnetic matching (GM) algorithm based on multi-layer constraint and K-nearest neighbor algorithm is proposed, and the fusion of ZUPT_HESO-pedestrian navigation algorithm and geomagnetic matching algorithm is realized, which improves the accuracy and reliability of pedestrian navigation. The actual data test proves that the proposed pedestrian navigation algorithm based on ZUPT_HESO_GM effectively improves the positioning accuracy by more than 79%. © 2017, Editorial Board of Acta Armamentarii. All right reserved.