Research On Indoor Positioning Based On Fusion Of WiFi Fingerprint And PDR

The development of mobile and Internet has made people's demand for Location Based Services increasingly strong.Although the Global Positioning System(GPS)performs well outdoor,it is difficult to receive GPS signals in indoor environment.Accurate and popularizable indoor positioning technology is still in research.Indoor positioning technology based on wireless signals relies on external signals and is vulnerable to fluctuations caused by environmental impact.Inertial navigation technology based on smartphones can provide relatively stable positioning results in a short time,but the error accumulates with time and produces serious drift.In this paper,we present a framework of combining smartphone sensors and WiFi for accurate indoor localization.The specific algorithms for WiFi fingerprint location,PDR location and fusion algorithm are designed respectively.To solve the problem of instability of WiFi signal,a fingerprint map construction method is proposed.We use the stability of Access Point(AP)and its signal strength as fingerprint features.In the offline states we proposed preprocessing and AP selection algorithms.In the online matching stage,the classical KNN algorithm is used as the prototype of the algorithm.Considering the influence of AP strength and stability on positioning,we calculated the feature distance based on AP weighting.Finally,we estimate the coordinates of position based on weighted Euclidean distance and physical distance together.Compared with the KNN algorithm,the proposed algorithm has increased the probability within 3 m of positioning error from 55.80% to 83.49%.Based on peak detection method,both the peak and valley are judged by threshold and time window to avoid false peak interference caused by equipment jitter in PDR positioning.A complementary fusion algorithm for solving the course angle is presented.The gyroscope's angle drift is corrected by filtering and smoothing the course angle of the compass,which is sensitive and high precision in a short period of time and the long-term stability of the compass.Finally,the angle drift caused by the gyroscope is corrected.After the algorithm is programmed and implemented,specific experiment shows the positioning trajectory of PDR module in this paper can basically reflect the real trajectory.In order to make full use of the complementary advantages of the two technologies,the extended Kalman filter is used to fuse WiFi fingerprint module and PDR module.The problem of frequency mismatch in multi-source information fusion is processed by time registration.Experiments are arranged to verify the effectiveness of the fusion positioning system.The positioning error in this paper is within 1.5m and the probability is 80%,and the positioning trajectory basically coincides with the real trajectory.