Research On Fingerprint Algorithm For Indoor Positioning Based On BP Neural Network

With the continuous improvement of people’s living standard,people’s demand for Location Based Service(LBS)are incressing rapidly.Outdoor positioning softwares like GPS and Beidou have brought us great convenience.In recent years,with the rapid development of Wireless-Fidelity(WiFi)technology,demands for indoor positioning service are also increasing.Focus of people’s research start to move from outdoor to indoor,and indoor positioning technology have obtained more and more attention.As outdoor technologies like GPS and BeiDou can hardly meet the requirements of complicated indoor positioning,more and more people start to study the indoor positioning of WiFi based on RSSI.In this paper,the traditional indoor positioning algorithm of WiFi based on RSSI has low positioning accuracy and large positioning error.In order to improve the positioning accuracy at indoor environment,an improved the indoor location algorithm based on BP neural network.Using BP neural network and fingerprint algorithm to improve the WiFi indoor location technology,Main contents and conclusions in this thesis are as follows:Firstly,analyzed and studied the research background of this topic,discussed the research dynamics of this topic at domestic and foreign,pointed out the direction of later studies of this thesis.Seconedly,summarized the relevant technical points and principles involved in this paper,such as indoor positioning fingerprint algorithm and BP neural network.And these methods were used to solve technical issues and difficulties exist in our research.Thirdly,described the related principles of the traditional indoor positioning fingerprint algorithm,analyzed data acquisition methords and processing methods of RSSI.And proposed a reasonable data collection program and data processing method,determined the Ranging Model Parameters of RSSI,designed the indoor positioning analysis system.Fourthly,analyzed the deficiencies of traditional BP neural network algorithms and proposed an improved BP neural network algorithm.Studied the indoor positioning fingerprint algorithm of BP neural network,and arranged the experimental environment with laptops to collect fingerprint data.Propossed RSSI data and use these data to construct BP neural network model.MATLAB was ued to simulate traditional indoor positioning fingerprint algorithm and BP neural network Algorithm,positioning errors of these two algorithms was compared to determine the merits of the algorithm,and simulation results verified the effectiveness of the algorithm.The simulation result is that the positioning mean square error of the BP neural network algorithm is 3.765,and the location error of the traditional algorithm is 6.62444.It shows that the BP neural network algorithm has small positioning error and high precision.