| With the development of Internet of things,computer network and other technologies,indoor positioning and navigation,logistics management,asset monitoring and other application scenarios require high-precision,high reliability and low-cost indoor positioning.As a low-cost,high-compatibility wireless technology,Bluetooth is widely used in various electronic devices.The traditional Bluetooth positioning method is based on Received Signal Strength Indication(RSSI),and there are still problems such as low positioning accuracy.In the Bluetooth 5.1 version,the Angle of Arrival(AOA)technology was introduced,which provided a new research idea for the realization of Bluetooth high-precision positioning.This thesis aims at the problem of high-precision positioning of fixed objects in indoor environment.Based on Bluetooth Low Energy(BLE),this thesis proposes a Weighted Fingerprint Feature Matching Location(WFFML)algorithm based on AOA and RSSI,and a Weighted Centroid Location based on AOA Triangulation(WCLAT).The corresponding positioning system is built,and the positioning results of the proposed algorithm are analyzed and verified through the collected actual data.The main work of this dissertation is as follows.(1)A weighted fingerprint feature matching algorithm based on AOA and RSSI is proposed.On the basis of traditional fingerprint positioning,the algorithm uses both AOA and RSSI value as matching features.In the offline stage,the natural breakpoint method is used to extract the fingerprint value,and the establishment of the fingerprint database is completed.Then the feature weight is trained through Random Forest.In the online stage,according to the proposed weighted threshold matching method,the weighted Euclidean distance is calculated by using the trained feature weight,and then the prediction coordinates of the target are obtained.After modifying and debugging the data collection code,the positioning system is built,and the positioning result of WFFML algorithm by using the data collected in the actual scene is analyzed and compared with the existing algorithms.The result shows that the WFFML algorithm is the best,the average error is only 42 cm,the accuracy within 0.5m error is 71%,and the accuracy within 1m error is 97%.(2)Weighted centroid localization algorithm based on AOA Triangulation is proposed.The principle of using IQ(In-phase/Quadrature)value to calculate AOA is studied.The algorithm simplifies the AOA calculation method of positioning platform,and optimizes its parameters by least square method.Then,based on triangulation method,a weighted centroid algorithm is proposed,which uses the power term of RSSI to weight multiple intersections to obtain the predicted coordinates.Similarly,after modifying and debugging the data collection code,the positioning is built,the system parameters are optimized and the positioning coordinates are calculated according to the actual collected IQ data.The result shows that the WCLAT algorithm reduces the calculation error of the AOA value by about 4°,and the average positioning error is about 0.62 m.Compared with the results of centroid algorithm before optimization,the average positioning error is reduced by about 0.5m.Although the positioning accuracy of WCLAT algorithm is not as good as WFFML algorithm,it has low implementation complexity,and the amount of data collection is less than 6% of WFFML algorithm. |
