Indoor Positioning Algorithm And Implementation Based On The Fusion Method Of WiFi And Bluetooth

With the development of mobile internet technology, Location Based Services (LBS) has been paid close attention by all the fields of the world. In LBS, the core technology is Location. At present, Global Positioning System (GPS) based outdoor positioning technology has been highly mature. On the contrary, for the indoor positioning field, although many popular technologies have been provided, there is no one can be used in any circumstance until now. The main reason lies in the complex indoor environment. The existing single mode positioning technologies can’t meet all the demands which increasing with people’s life, and their limitation has become increasingly prominent. In recent years, researchers try to fusion multiple technologies. The focus of the study is that how to achieve the complementary through the advantages and disadvantages of the various technologies.This paper analyses all kinds of typical indoor positioning technologies and existing fusion positioning systems in order to clarify that, WiFi and Bluetooth has good prospect to combine together and make one fusion positioning technology. The paper puts forward one kind of indoor positioning fusion scheme based on WiFi and Bluetooth in order to improve the positioning accuracy. The scheme is based on the Location Fingerprint Algorithm. As WiFi signal has wide range, huge amount of data and unstable RSS, this paper proposes a location fingerprinting algorithm based on KL divergence kernel function for WiFi positioning stage. It can capture the non-linear model of WiFi RSS. As Bluetooth signal has short range and more stable RSS than WiFi, this paper proposes an improved fingerprint algorithm for Bluetooth. It can retain the positioning information in original Bluetooth signal data so that reduce the error caused by the loss of information in data filtering. First two analysis stages provide data for the final fusion scheme as intact as possible. In the fusion positioning stage, this paper presents a decision layer fusion model based on K-means clustering algorithm. It uses the data from Bluetooth to modify the positioning result analysed from WiFi, in order to make the positioning result closer to the true value.In this paper, the content about the experiment describes the whole process of measurement and analysis in the the real environment. It uses Andriod mobile phone as the receiving terminal to collect the data. It also solves the problem of WiFi and Bluetooth signals synchronization. Finally we implement the algorithm by Matlab and get the location result. The experiment shows that, by using fusion positioning algorithm provided by the paper, the error is significantly reduced compared to the single mode WiFi positioning method. Compared with the general fusion method of WiFi and Bluetooth, the positioning effect is also improved obviously.