Research On Indoor Location Algorithm And Design Of Location System Based On WIFI/PDR Fusion

With the growing demand for indoor location services,indoor location technology has become a research hotspot in recent years.At present,with the development of smartphones and the Internet,various inertial sensors are widely integrated with smartphones,and WIFI facilities are almost fully covered in many buildings.Therefore,the WIFI fingerprint positioning technology and PDR(Pedestrian Dead Reckoning)positioning technology are deeply analyzed based on the smartphone positioning platform in this thesis,and the WIFI/PDR fusion technology is studied by combining the advantages of both.In the research of WIFI fingerprint location technology,Based on the phenomenon of non-uniform spatial resolution of RSSI(Received Signal Strength Indicator)signal strength,a new weight factor was introduced into the European distance calculation method to improve the accuracy of similarity calculation.In the WIFI fingerprint location stage,the traditional WKNN(Weighted K Nearest Neighbor)algorithm only takes the signal strength similarity as the evaluation indicator,which leads to the phenomenon that the signal strength of the reference point and the observation point is highly similar and the real position is too deviated.To solve this problem,this thesis proposes a WKNN algorithm based on a sliding k-window.Firstly,the position based on the traditional WKNN algorithm is estimated,Then,the position similarity between reference points is improved by detecting the dispersion between the position of each reference point and the estimated position.The experimental results show that the proposed matching algorithm can improve positioning accuracy.In the research of PDR positioning technology,it is studied from three directions: step frequency detection algorithm,step estimation model,and heading angle solution.In the research of step frequency detection,the combined acceleration strategies under different postures are subdivided;In the research of step size estimation model,variance and step frequency are introduced into the traditional Weinberg step size model to improve the positioning accuracy to a certain extent;In the course of course angle calculation,Kalman algorithm is used to fuse the two methods of magnetometer and gyroscope to improve the accuracy of the solution.In the research of extended Kalman filter fusion algorithm,it is found that the WIFI location results of jump and aggregation will also make the location trajectory of WIFI / PDR fusion jump and aggregation.Taking advantage of the higher displacement accuracy of PDR in a short time,this thesis proposes an extended Kalman filter algorithm based on PDR displacement constraints.Firstly,the algorithm filters the reference points in the WIFI positioning stage based on the predicted position of the fusion algorithm,and then further selects the WIFI Positioning results based on PDR displacement constraints.The experimental results show that the fusion algorithm in this thesis is better than the traditional extended Kalman filter fusion algorithm,The positioning result is better,and the average positioning error decreases significantly.Finally,based on the above research content,this thesis develops the android system as the client,uses the Ali Cloud server to build the My SQL database and the back-end system,and completes the design and implementation of the WIFI/PDR integrated indoor positioning system.Experimental results show that the extended Kalman algorithm based on PDR displacement constraint can reduce the jumping WIFI location results to a certain extent and improve the location accuracy.