Research Of Wireless Indoor Localization With The Bluetooth Module

Location based service has been widely used in outdoors with the maturity of GPS technology. However, GPS cannot meet the accuracy and efficiency requirements of indoor localization, thus as the core technology of indoor location based service, indoor localization has become a hot research direction. Among the wireless communication technologies that are applied to indoor localization, BLE (Bluetooth Low Energy) with low power consumption, long distance transmission and low latency characteristics has been widely concerned in the field of indoor localization.This thesis studied indoor localization with the Bluetooth module. Two localization methods are proposed with the range-based and fingerprinting technique, moreover, we developed an indoor localization algorithm verification system with the Bluetooth model. The main work includes:1. Based on distance measurement, we proposed an indoor localization method which achieved to selected best APs and applied enhanced weighted localization algorithm to estimate location. The method is defined as follows: firstly, fit the function between distance and RSSI by log-distance path loss model, calculate the distance with RSSI. Secondly,to select the best APs, the method will cluster the circles. The APs locate at the center of the circles with the propagation distance being the radius.The cluster with most circles constitute the best APs. Thirdly, the enhanced weighted trilateration localization algorithm uses positioning unit quality as the weighting factor to estimate the position. Enhanced weighted trilateration localization algorithm with best APs has a high precision, it improves the localization accuracy by 1.76m compare to traditional trilateral centroid localization.2. Based on fingerprinting technique, we proposed an indoor localization method which established the offline fingerprint database and estimate location with IW-KNN, the method is defined as follows: firstly,dealing the offline fingerprint data by using Canopy clustering algorithm and the k-means clustering algorithm. Secondly, IW-KNN algorithm improves the calculation of similarity between RSSI and adds weighting factor to do the iterative calculation. Clustering the fingerprint data can reduce the time of database queries effectively. Moreover, IW-KNN improves the localization accuracy by 2.12m compare to traditional KNN localization method.3. The indoor localization method verification system includes the data collect APP, the APP implements the communication of BLE, data collection and user management. Intelligent terminal is used to collect RSSI in the indoor environment which is deployed with BLE devices.Positioning experiments are conducted with RSSI, the results are given to verify the localization methods we proposed.