Research On Indoor Localization Method Based On Received Signal Strength

The indoor environment is a primary place for human work and life.With the development of Internet and communication technologies,indoor localization technology is applied in more and more services.Indoor localization method is the key point to the indoor localization technology.The indoor localization method based on the received signal strength has many advantages such as low hardware resource requirements,simple data acquisition mode,and abundant data processing algorithms.It has become a hot spot in indoor localization methods.In this dissertation,several key methods of indoor localization based on received signal strength are deeply studied,including: the localization algorithm based on geometric principle and the indoor localization method using fingerprint database,especially in the aspects of rational use of anchor point,fingerprint database update and matching algorithm.The main contributions and innovations of the dissertation are as follows:1.Based on the analysis of indoor wireless signal propagation and channel characteristics,the dissertation focuses on the extension of geometric localization methods in outdoor environment.According to the advantage of plenitudinous anchor points in indoor environment,a bilateral greedy iterative localization method is proposed.Based on two anchor points,it gradually introduces each effective anchor point,thus making full use of the position information of multiple anchor points in the room.The improved method of this dissertation is tested in the indoor environment covered by Wi-Fi network.The test results show that compared with the ordinary trilateral and fingerprint localization method using a certain density reference point,without significantly increasing the amount of calculation,the bilateral greedy iterative localization method improves the localization accuracy.2.The nearest neighbor matching algorithm based on fingerprint database tends to focus only on the amplitude of the received signal,while ignoring the positional directivity.This dissertation proposes an improved nearest neighbor matching algorithm based on the mean correlation coefficient.The proposed algorithm can not only match the received signal strength value,but also analyze the correlation between the positioning target and the reference point,thus improving the matching accuracy.In the real indoor floor,the experimental test environment was built.According to the actual collected reference point fingerprint data,the fingerprint database was established,and the target test point was selected to test the algorithm.The test results show that compared with the traditional absolute nearest neighbor algorithm and the correlation coefficient method,the improved algorithm achieves an outstanding matching effect without adding extra computation.3.Indoor real-time localization technology generally requires inertial measurement unit,however,inertia unit has error drift,and many wireless devices in indoor localzation applications are not equipped with inertial measurement unit.Based on nearest neighbor matching algorithm and mean correlation coefficient,this dissertation proposes a real-time localization algorithm combined extended Kalman filter.Firstly,the improved matching algorithm is used to achieve the target location,and then the extended Kalman filter algorithm is used for real-time updating,thus realizing the indoor real-time localization.In the indoor environment covered by Wi-Fi network,the proposed algorithm is tested.While the result shows that compared with the traditional point-by-point localization scheme,the real-time localization algorithm of this dissertation has higher localization accuracy without matching each point and owns a better real-time performance.4.In the traditional fingerprint-based indoor localization method,establishing and maintaining the fingerprint database is an vital task.Generally,the original fingerprint database needs to be established in a long period of time,and the fingerprint database must be updated in time due to the variability of the indoor environment.In this dissertation,a fingerprint updating system is designed and Gaussian process regression algorithm is used to implement database update on extended Kalman filter.Meanwhile,in the localization matching stage,combined with the stepwise subset localization strategy,and the modified cosine similarity method is used to achieve better matching.The test results of the real indoor environment show that the proposed fingerprint update system combined with the subset localization can greatly improve the positioning accuracy and significantly reduce the amount of calculation.