A Novel Dimension Reduction Algorithm for Fingerprint Positioning based on GrDOP and Geometric Constraints

Wen Liu, Xinyu Zheng, Zhongliang Deng, Lianming Xu

Abstract:	In fingerprint positioning, in addition to the higher positioning accuracy of high-dimensional signal feature (SF) fingerprints, there is a problem of large amount of calculation in the positioning process. Aim at the problem above, we define a parameter to evaluate the contribution of access points to positioning: Gradient Dilution of Precision(GrDOP), and we proposed a novel offline dimension-reduce algorithm based on GrDOP and Geometric Constraints. Experiments have shown that several positions on the map that are far away from each other are likely to have extremely similar signal characteristics. This common phenomenon will lead to large fluctuations in positioning results during the online phase. Aiming at issue above, we proposed a GrDOP Geometric Constraints Weighted k-nearest neighbors(GrDOP-GC-WKNN) online matching algorithm. In the proposed algorithm, the method of dual adaptive clustering is used to classify reference points (RP) in signal space and geometric space, calculate the GrDOP in every clusters and reduce dimension according to GrDOP; In the GrDOP-GC-WKNN online positioning algorithm, the positioning result of the previous state is used to position the current state positioning result to reduce the positioning error. An experiment was carried out at the second floor of the fourth building in Beijing University of Posts and Telecommunications. The positioning accuracy of the algorithm reached 1.22 meters in 80% of cases and real-time increased 2400% compared with the traditional fingerprint positioning. The algorithm we proposed provides a theoretical method for indoor high-precision real-time positioning, provide an accurate indoor solution to assist GNSS for accurate seamless positioning. Key words: Fingerprint position, GrDOP, Dimensionality reduction, Geometric constraint.
Published in:	Proceedings of the 31st International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2018) September 24 - 28, 2018 Hyatt Regency Miami Miami, Florida
Pages:	666 - 673
Cite this article:	Liu, Wen, Zheng, Xinyu, Deng, Zhongliang, Xu, Lianming, "A Novel Dimension Reduction Algorithm for Fingerprint Positioning based on GrDOP and Geometric Constraints," Proceedings of the 31st International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2018), Miami, Florida, September 2018, pp. 666-673. https://doi.org/10.33012/2018.16044
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In