Indoor localization method based on pedestrian dead reckoning aided by multi-source fusion

In view that traditional PDR localization technology has serious error accumulation problem, and the accumulated error is caused by course deviation, an improved course estimation solution is proposed by using geometric information to compensate and enhance the pedestrian course. We make such analysis that the exogenous absolute location is used to improve the positioning result of pedestrian dead reckoning (PDR), then an evolved EKF filtering algorithm with adaptive model noise is proposed, which realizes the PDR and WIFI positioning sources' filtering fusion. Experiment results show that, compared with traditional PDR, the PDR with enhanced-course can effectively restrain the error accumulation, and the total errors are controlled to about 5 m. The accuracy of original-course PDR adaptive filtering with WIFI location is improved by 82.8%, compared with the pure original-course PDR's one. And when fusion with enhanced-course PDR and WIFI location, the accuracies are improved by 90.2% and 49.5% respectively compared with the pure original-course PDR's and enhanced-course PDR's. These results show that enhancing course and fusion with exogenous absolute position can both effectively restrain the error accumulation. Considering the constraint conditions, the improved-course PDR schemes are applicable to a regular indoor environment. However, the original-course PDR and WIFI fusion scheme is not affected by the indoor structure, and its results with couple filtering calculation can meet the accuracy requirements for pedestrian indoor positioning. © 2016, Editorial Department of Journal of Chinese Inertial Technology. All right reserved.