| With the rapid increase of the demand for location in indoor scenes,the indoor location technology based on Bluetooth fingerprint has become one of the research hotspots of wireless location technology with the advantages of low cost,easy deployment and universality.However,due to the complex indoor environment structure,noise interference and other phenomena,the positioning performance of Bluetooth is seriously affected.At the same time,Most of the Bluetooth location methods using location fingerprint are time-consuming and labor-consuming,and the overall positioning accuracy is not high.Based on the shortcomings of the existing Bluetooth fingerprint positioning technology,it is of great significance to carry out the research on the location fingerprint positioning technology based on Bluetooth to promote the development of indoor location service industry.This topic takes the intelligent ROS mobile car as the research carrier,takes the position fingerprint localization as the theoretical basis,and makes multiple improvements from the offline stage and online stage of Bluetooth fingerprint positioning respectively,so as to improve the overall positioning accuracy of Bluetooth fingerprint in the indoor.This paper mainly completed the following research work.(1)In order to solve the problem of heavy workload of fingerprint reference point collection in off-line phase,an optimized method based on interpolation method to quickly build fingerprint database was proposed.Firstly,aiming at the dynamic change of Bluetooth signal,the second filtering algorithm was used to filter RSSI data effectively;then,in order to avoid the disadvantages of the traditional logarithmic path loss model,the model was modified and a segmented path loss model based on sliding window was proposed;finally,the interpolation method was used to construct the fingerprint database,and the fingerprint information of interpolation points was calculated based on the segmented path loss model,which could ensure the accuracy of the fingerprint database,reduce the offline sampling workload,and improve the efficiency of fingerprint database construction.(2)In order to reduce the amount of matching computation in online fingerprint matching,K-means clustering algorithm was designed to divide and cluster the constructed fingerprint database.As for how to divide clusters correctly,the elbow method based on slope judgment was used to determine the optimal cluster number k;At the same time,aiming at the selection of initial cluster center,a selection strategy of initial cluster center was designed to reduce the number of iterations.The experimental results showed that the optimized K-means clustering algorithm could obtain better clustering results.(3)In view of the fact that directly ignoring the loss of Bluetooth signal would lead to the increase of positioning error,this paper dealt with this kind of phenomenon in the separate line stage and the online stage.At the same time,the online phase focused on the research of WKNN location matching algorithm.Aiming at the problem that fixed K value was easy to produce error,an improved dynamic K value WKNN algorithm was designed,so that different environments could choose the better K value.To a certain extent,the positioning performance of WKNN algorithm was improved,and the overall positioning accuracy was improved.Finally,the mobile car was tested,and the results showed that,compared with the traditional Bluetooth fingerprint positioning accuracy,the multi optimized Bluetooth fingerprint scheme designed in this paper could effectively improve the positioning accuracy,so the simple fusion of Bluetooth fingerprint plus INS could better improve the positioning accuracy of indoor mobile car. |
PDF Full Text Request