|
Different positioning methods currently have limitations in respective application scenarios, so the integration of positioning has become a hot topic. In terms of data integration, inconsistent and fluctuations of the positioning data makes it difficult when fit the data. To solve this problem, the paper presents a method for position data filtering and data fusion normalization. Firstly, heterogeneous data sets are filtered with R-EWMA algorithm to obtain higher quality fusional data, and then fusion data are normalized with a modified Z-score normalization method without affecting the positioning accuracy and confidence. The study is mainly targeted at preprocessing work of multi-source heterogeneous data fusion, and to some extent excluding the invalid data. Experimental results show that TDOA and RSSI Location noise fluctuations have improved significantly through the methods, and the normalized space is in accordance with the fusion requirement, which will be effectively applied to the next study of fusion location. |
|
Keywords:Wireless Communication Technology; Guidance, Navigation and Control; Data Fusion; Indoor Positioning; R-EWMA Algorithm; Normalization |
|