Electronic Science and Technology ›› 2020, Vol. 33 ›› Issue (8): 74-79.doi: 10.16180/j.cnki.issn1007-7820.2020.08.013

Previous Articles     Next Articles

Indoor Positioning Algorithm Based on WiFi-BP

ZHU Yifeng   

  1. Director’s Office,Shanghai Engineering Research Center of Agriculture IOT,Shanghai 200335,China
  • Received:2019-05-23 Online:2020-08-15 Published:2020-08-24
  • Supported by:
    Shanghai Key Project of Science and Technology Agriculture (Hu Nongke Attacking Words (2014) 7-4-1)


Aiming at the problem that signal deviation caused by the device variability affects positioning accuracy, an indoor positioning algorithm combining BP neural network and weighted centroid positioning algorithm was proposed. In this paper, the RSSI data of different mobile phones were cleaned by outlier detection algorithm, and the cleaned data was used as the data source of BP neural network to train the model, thus obtaining a stable nonlinear BP model. On this basis, combined with the improved indoor positioning algorithm for indoor positioning. Experiment results showed the mean error, minimum error and maximum error of the proposed algorithm were 0.58 m, 0.24 m and 1.06 m, respectively, and the positioning accuracy was significantly higher than that of the existing similar algorithms.

Key words: the device variability, received signal strength indication, indoorpositioning, BP neural network, the outlier detection, weighted centroid positioning

CLC Number: 

  • TN92