Research On The Fingerprint Positioning Technology Based On Low-energy Bluetooth

With the popularity of smart mobiles and the rapid development of Internet technology,location based service gets more and more attention of people and it owns great potential in medical treatment and health,safety and information forwarding.Obtaining users' location information is the basis of enjoying the local based service.Nowadays people work and live indoors most of time and the GPS positioning system can not adapt to the complex environment indoors,so it becomes the focus of current research that using wireless technology to realize the indoor positioning.Due to its high power consumption,bad penetrability and short signal transmission distance,the traidtional bluetooth technology was not widely promoted.However,the release of low-energy bluetooth technology makes it possible to use it in indoor positioning.Because of its improvement in power consumption,transmission distance and signal strength,this paper studies the indoor localization algorithm on the base of the low-energy bluetooth technology.This paper analyses the indoor distribution characteristics of bluetooth signal and the factors that influence distribution characteristics. It also tests the degree of affecting signal strength by measuring distance,transmission path and equipments.Because of the complicated indoor environment,the signal propagation model method produces large error.The fingerprint positioning method is more applicable to the indoor environment.However,this traditional method has obvious shortcomings in time complexity and accuracy. So in view of the defect of it, the improved measures are taken in this paper.K-means clustering and fuzzy c-means clustering methods are used to analyse the initial fingerprint database which is divided into several subclasses.Real-time data matches the fingerprint data in a certain subclass instead of the whole database.This reduces the search space and thematching time and weakens the influence of fingerprint database.Moreover,the original weight coefficeint is improved to enhance the accuracy of boundary point and the point estimation in the class will not be affected.At last,the simulation is completed by matlab after the low-energy bluetooth signals are collected.The FWKNN algorithm based on the fuzzy c-means clustering can reach 80% in the range of one meter error.The result shows that the improved fingerprint algorithm enhances the precision of positioning effectively compared with the traditional algorithm.