Research On Positioning Technology Based On Joint Filter Fusion Prediction

With the continuous improvement of science and technology and living standard,people's requirements for location service also increase.At present,with the continuous development of satellite technology,GPS,Beidou and other satellite positioning systems can basically provide users with good outdoor services.Due to the complexity of indoor environment,wireless signals are greatly interfered,which makes indoor positioning based on these signals cannot always provide high quality location service.How to obtain a more accurate positioning has become the goal of researchers.This thesis hopes to improve the accuracy of indoor positioning by studying the positioning technology based on joint filtering fusion position prediction.The main work is as follows:1.Aiming at the problem that a stable and reliable location fingerprint database cannot be constructed in the process of WiFi fingerprint positioning,this thesis proposes a joint filtering algorithm based on Kalman filtering and neighborhood mean filtering.The fingerprint database is filtered before and after the construction to ensure the reliability of the location fingerprint database.Clustering algorithm and Lagrange interpolation method based on window were introduced to improve Kalman filter to reduce the influence of noise in the process of collection,and the fingerprint database was constructed.Furthermore,the fingerprint database is optimized by the improved neighborhood mean filtering algorithm.In the process,the weight of non-noisy points in the target neighborhood is fully considered,and different weights are given to the target points according to the Euclidian distance between the non-noisy reference points and the target points and the standard deviation of the reference points to influence the value of the target points.The experimental results show that the WIFI positioning error control based on joint filtering proposed in this thesis accounts for 37% within 1m,and the probability of error within 2m is 70%.2.Aiming at the problem that the accuracy and robustness of a single positioning technology may degrade in a complex indoor environment,this thesis proposes a positioning method that integrates WiFi and location prediction.The user's historical trajectory is analyzed and the Markov location prediction model is constructed.A method to judge the reliability of WiFi location results was proposed,and a fusion method based on the previous time position constraint was proposed to improve the positioning accuracy.The location prediction results were fused to improve the location accuracy,and the optimization algorithm based on gradient approximation was used to further optimize the fusion location.The experimental results show that the fusion positioning method proposed in this thesis has a probability of error within 1m of 45%,and an error within2 m of 79%,and the positioning error can be controlled within 4m.Compared with the existing positioning methods,the proposed method has higher positioning accuracy and positioning stability.