Research On Tracking Method Of Indoor Personnel Based On CSI And PDR

Location based service is playing an increasingly important role with its unique charm both indoors and outdoors.When describing human behavior habits,human walking trajectories are more complicated than a single location.Therefore,the tracking based service presented to users based on location and trajectory is more refined than LBS,and is more favored by scholars.With the development of technology,more and more indoor location services are implemented on smart terminals.Wi-Fi positioning technology based on smart terminals has gradually become an important development direction for indoor location services.However,a single indoor trajectory tracking method usually has limitations in indoor location tracking,and the joint positioning and tracking of different methods can effectively improve the tracking accuracy.In order to provide users with better indoor trajectory tracking services,this article conducts in-depth discussion and research on trajectory tracking algorithms,equipment costs and tracking accuracy.The main work of this thesis are as follows:(1)In view of the high complexity and high computational cost of current indoor pedestrian trajectory tracking methods,this research proposes an indoor pedestrian trajectory tracking method based on K-Nearest Neighbor.First collect fine-grained CSI data,perform feature extraction on the collected data,and perform filtering to establish an offline fingerprint database;In the online phase,the real-time data and the data in the offline database are matched using the KNN classification algorithm position fingerprint algorithm fused with Kalman filter to realize the trajectory tracking of the target person.(2)Aiming at the accuracy of indoor pedestrian trajectory tracking,an indoor pedestrian trajectory tracking method based on Hidden Markov Model is proposed.First,according to the CSI data collected by the target walking trajectory,determine the time when the target starts and ends.Then determine the direction of movement according to the velocity direction estimation model,Then according to the speed estimation model,the movement speed is determined and corrected so as to obtain an accurate speed estimation.Finally,the hidden Markov model is combined to realize the real-time tracking of pedestrians.(3)The continuity of indoor people's movement leads to difficulties in trajectory tracking,a dual-mode signal fusion method is proposed.The sensor signal in the smart phone is fused with the CSI signal collected by Wi-Fi to obtain a more accurate tracking result.Accurate initial position is a key step in indoor trajectory tracking.Firstly,the accurate initial position is obtained by data preprocessing through the CSI signal collected by Wi-Fi.Then combined with the pedestrian dead reckoning algorithm to achieve continuous trajectory tracking of pedestrians.