DENG Su, XUE Feng, YU Min. WiFi fingerprint indoor location method with BP neural network based on improved artificial fish swarm optimization algorithm[J]. GNSS World of China, 2020, 45(1): 82-87. DOI: DOI:10.13442/j.gnss.1008-9268.2020.01.014
Citation: DENG Su, XUE Feng, YU Min. WiFi fingerprint indoor location method with BP neural network based on improved artificial fish swarm optimization algorithm[J]. GNSS World of China, 2020, 45(1): 82-87. DOI: DOI:10.13442/j.gnss.1008-9268.2020.01.014

WiFi fingerprint indoor location method with BP neural network based on improved artificial fish swarm optimization algorithm

  • In view of the traditional indoor localization algorithm based on BP neural network existing low precision and sconvergence speed. considering the sophisticated indoor environment, there is usually the multipath effect,in addition that the signal attenuation model unsuitable for accurate positioning, this paper proposes an improved artificial fish optimization WiFi fingerprint indoor localization algorithm of BP neural network. the foraging and searching methods of artificial fish are used to improve the speed and ability of global optimization, using the improved artificial fish algorithm (IAFSA) to optimize the selection of weights and thresholds of BP neural network, which effectively avoid the disadvantage that predicted value of traditional BP neural network is easily plunged into partial optimum, the signal is denoised by gaussian filter in advance.At the same time, the relationship between the signal strength value (RSSI) obtained by the sampling point and the position coordinate is established. Experimental results show that compared with the traditional BP neural network method, the proposed method of this paper reduces the average positioning error by 0.75m, and the average positioning accuracy is improved by about 32.2%. The algorithm of this paper improves the reliability of positioning and has better stability.
  • loading

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return