Research And Implementation Of Indoor Positioning Method Based On Location Fingerprint

Positioning related services are widely used in rescue,navigation and other scenes.Indoor positioning,as the basis and key of positioning field,has important research significance.Among them,the WiFi indoor location method based on location fingerprint is easy to popularize and practical.Therefore,this paper takes this as the research direction to improve the existing shortcomings.Firstly,taking improving the positioning accuracy as the starting point,the problem that the weighted k-nearest neighbor(WKNN)algorithm assigns the same weight to each WiFi signal source(Access point,AP)is solved.Each AP is weighted,and on this basis,according to the mathematical relationship that the original weight and the number of items are reciprocal to each other,an equation transformation of distinguishing weight is proposed,which can distinguish weight more effectively.The experimental results show that the positioning accuracy of distinguishing AP weighted Euclidean distance WKNN algorithm is better than AP weighted Euclidean distance WKNN algorithm and traditional Euclidean distance WKNN algorithm.Then,in terms of optimizing location efficiency,this paper applies KD tree to location fingerprint method,and improves the pre ranking method to further reduce the complexity of the algorithm.The experimental results show that the location fingerprint method of the improved KD tree is more efficient and real-time than the location fingerprint method of the unmodified KD tree and the traditional location fingerprint method.In addition,the space near the stair corner is narrow and far from AP,resulting in poor positioning accuracy in this area.To solve this problem,this paper uses the Bluetooth+WiFi fusion location algorithm based on location fingerprint in this area to fuse the Bluetooth estimated coordinates and WiFi estimated coordinates at the decision-making level.Other areas still use the WiFi location algorithm based on location fingerprint,which not only improves the location accuracy but also does not consume too much cost.The experimental results show that the Bluetooth+ WiFi fusion location algorithm can improve the local location accuracy in the narrow area near the corner of the stairs.Finally,in the realization of three-dimensional positioning,this paper uses the fully connected condensed hierarchical clustering algorithm for floor discrimination.This method has high accuracy,strong real-time performance,and does not need the help of additional hardware equipment.The experimental results show that the floor discrimination method based on fully connected condensed hierarchical clustering has more advantages in discrimination accuracy and real-time positioning than other floor discrimination methods.Based on the above theoretical research,this paper designs and implements the indoor three-dimensional positioning application software,which is mainly divided into three modules:Wechat Mini Program,server and SQLite3 database.It realizes the functions of data acquisition and three-dimensional position display.It has low dependence on hardware devices,strong scalability and is suitable for large-scale promotion.