Research Of Manifold Learning Based Bluetooth Positioning Algorithm

The proposed BLE technique has attracted wide attention of researchers in indoor positioning.The BLE technique is helpful to the development of indoor positioning technology because of the property of low power consumption,low cost,easy to layout,and so on.As wireless sensor,the signal of Beacon is easy to be disturbed in the indoor environment,thereby reducing the positioning accuracy.Therefore,in order to satisfy the people's high precision positioning demand,this paper study the method of indoor positioning.Based on the analysis of the existing indoor positioning technology and indoor positioning algorithm,this paper focuses on the deterministic localization algorithm of fingerprint localization algorithm.The main work of this paper is as shown below.1.In view of the particularity of the indoor environment,the signal feature of Beacon on different transmitting frequency is studied.According to the analysis of the signal characteristics,the Kalman filter is used,and this paper propose the filter method of keeping the maximal RSSI.2.See the indoor positioning as the problem of high-dimensional nonlinearclassification.The positioning algorithm based on fingerprint can response effectively to the complex indoor environment.Manifold learning finds the real relationship between data by finding the low-dimensional embedding for high-dimensional data.Therefore,to improve the localization accuracy,this paper measures the distance of RSS vectors on their low-dimensional manifold and proposes a novel positioning method IWKNN.3.We layout those Beacon uniformly.The main experiments are as follows:(1)We compare the performance of WKNN/EWKNN and their ISOMAP enhanced versions.And then,we analyze the advantages of using ISOMAP for indoor positioning.(2)In order to test the performance of proposed method,we compare the proposed IWKNN with Trilateration?IoT?Bayes?GMM?wKNN-Bayes?WKNN.It verifies the effectiveness of IWKNN.