The use of smartphones for indoor positioning has become increasingly popular in recent years. The cumulative error is an unavoidable problem for pedestrian dead reckoning (PDR). Mismatching is the main problem for fingerprint matching. Therefore, how to integrate multiple sensors to reduce the position error and improve the robustness of the positioning system is a subject worthy of further research. For fingerprint matching, an enhanced dynamic time warping (EDTW) is proposed to improve the accuracy of magnetic field matching. For PDR/magnetic matching (MM)/WiFi, a multi-source fusion positioning approach based on a robust extended Kalman filter (REKF) that introduces the estimation of the innovation sequence covariance is studied. The PDR-based error model is taken as the state transition equation and the difference between PDR and MM and the difference between PDR and WiFi as the observation equation. The experimental results show that the multi-source fusion positioning approach not only reduces the position error but also improves the robustness.

近年来，使用智能手机进行室内定位已变得越来越流行。对于行人航位推算（PDR），累积误差是不可避免的问题。不匹配是指纹匹配的主要问题。因此，如何集成多个传感器以减少位置误差并提高定位系统的鲁棒性是值得进一步研究的课题。对于指纹匹配，提出了一种增强的动态时间规整（EDTW）以提高磁场匹配的准确性。对于PDR /电磁匹配（MM）/ WiFi，研究了一种基于鲁棒扩展卡尔曼滤波器（REKF）的多源融合定位方法，该方法引入了创新序列协方差的估计。将基于PDR的误差模型作为状态转换方程，将PDR与MM之间的差以及PDR与WiFi之间的差作为观察方程。实验结果表明，多源融合定位方法不仅减小了位置误差，而且提高了鲁棒性。