| At present,with the rapid development of Internet technology and smart mobile devices and the emergence of Bluetooth low energy consumption,Bluetooth-based indoor positioning technology has attracted more and more attention from academia and related industries.Although the fingerprint indoor positioning method based on Bluetooth has been continuously improved and perfected in recent years,there are still some problems that need to be resolved:In the natural environment,the Received Signal Strength Indication(RSSI)does not obey the normal distribution.The quality of RSSI fingerprints obtained by conventional filtering methods is low;Weight K-Nearest Neighborhood(WKNN)has defects in selecting adjacent reference points,resulting in low positioning accuracy.An indoor positioning method based on distance fingerprints was designed in this thesis to solve the above problems.An RSSI-range model based on RSSI sequences and BackPropagation neural network by analyzing the indoor RSSI distribution and the time-domain variation characteristics of RSSI was designed.Then,the RSSI fingerprints were converted to the distance fingerprints according to the RSSI-range model,and the distance fingerprint database was established for matching fingerprints.Finally,a location area selection algorithm and an adaptive K value algorithm was proposed to achieve the improvement of the WKNN algorithm.By narrowing the reference point search space and dynamically changing the K value to eliminate reference points that interfere with positioning,the positioning accuracy was improved.The experimental results show that the RSSI-range model is better than the least-square fitting model in the accuracy and fitting the relationship between RSSI and distance.The average positioning error of the indoor positioning algorithm based on distance fingerprints is0.86 m.Compared with the traditional positioning method based on fingerprint,the positioning accuracy is improved by 33.29%. |
