Research On Indoor Localization Method Based On Location Fingerprint Database

With the continuous iteration of science and technology and the continuous improvement of people’s economic level,people’s lives will gradually become interconnected today and in the future.There will be more and more smart devices in the indoor environment.This also makes indoor location services increasingly important.WIFI,as a mature technology,has been widely used indoors,and fingerprint positioning technology has a good prospect for its advantages of convenient deployment,low cost and non-line of Sight(NLOS)influence.The layout method of traditional base station Access Point(AP)is usually to randomly select the location access or arrange as many AP as possible,which will increase the technology landing cost,but also does not consider the impact of the main indoor people.The traditional methods of artificial measurement and calibration of coordinate system will introduce errors,resulting in the final positioning results do not match the actual environment.At the same time,the change of wireless signal distribution in indoor environment will cause a large deviation between the fingerprint database collected in offline stage and the data collected in online matching stage,thus making the estimation results of target points inaccurate.In order to solve the problems of cost and error in traditional AP location selection,a base station layout algorithm based on particle swarm optimization is proposed in this thesis.The algorithm integrates the thought of analytic hierarchy to consider cost and effective coverage,and finally can find the optimal AP location through population iteration.At the same time,aiming at the coordinate calibration problem of AP points,this thesis proposes an algorithm based on Extended Kalman Filter(EKF)for AP point coordinate self-calibration.This algorithm can realize the establishment of coordinate system and coordinate self-calibration after the location of base station is selected with less labor and time cost.The ultimate prevention proved to reduce both cost and error.This thesis focuses on the two key points of labor time cost and location fingerprint quality in database construction,as well as the limitations of fingerprints collected in the offline stage,and puts forward a fingerprint database interpolation algorithm based on time-varying collaborative Kringing.Firstly,the characteristics of indoor WIFI RSSI signal are analyzed.Compared with the traditional algorithm,the time dimension considering human factor is added,and the influence of human body and distance on signal is expressed by exponential and Gaussian variogram models respectively.Finally,simulation experiments prove the effectiveness of the algorithm,and it can reduce the labor and time cost in offline stage.