Research And Implementation Of Multi-source Indoor Positioning Information Fusion Technology

With the deepening of the information revolution,new technologies such as the Internet of Things and the artificial intelligence are changing our life.Location service is an indispensable part of technology innovation.The indoor positioning based on wireless LAN has lower cost and better accuracy.And it is the most promising indoor positioning method for large-scale application.In addition,in recent years,researchers have studied the positioning method using a terminal sensor to measure the pedestrian's dead-reckoning position.It can accurately calculate the user's position within a short distance.But the cumulative error occurs over a long period of time.Due to complementarity of positioning information,multi-source positioning information fusion is an effective way to improve positioning accuracy.The combination of pedestrians' dead reckoning and WIFI positioning results is the most popular research direction.However,in the case of high personnel density,the WIFI signal is unstable and affects the positioning accuracy.This paper focuses on the multi-source indoor positioning information fusion technology based on particle filter.Combining acoustic distance measurement technology,WIFI positioning technology and pedestrians' position estimation technology,a multi-user joint particle filter algorithm is proposed to fuse positioning information and improve indoor positioning accuracy.Aiming at the problem of mutual interference of acoustic ranging in the process of multi-user positioning,a multi-user adaptive matching algorithm was proposed to dynamically allocate the user's ranging object and the used acoustic band to avoid the same frequency interference.In view of the lack of terminal computing power,this paper studies and designs an indoor positioning system based on WIFI probes.First,the characteristics of multi-user multi-source indoor positioning information are analyzed and analyzed.The positioning method based on the WIFI fingerprint database obtains the positioning results independently in each positioning process.But the positioning accuracy is insufficient and greatly affected by the flow of people and other factors in the environment.Pedestrian dead reckoning can obtain more accurate positioning results in a short period of time.But cumulative errors will occur for a long time.The fusion method based on positioning information is generally particle filtering and Kalman filtering.The particle filter can work in non-linear systems and non-Gaussian noise.And it is more suitable for indoor positioning systems.Next,the user's joint particle filter algorithm proposed in this paper is described in detail.The algorithm integrates the WIFI positioning results,pedestrians' heading measurement results,and acoustic distance measurement results in a joint particle filter to improve positioning accuracy.In order to solve the problem that multiple users interfere with acoustic ranging in the same area,this paper proposes a multi-user adaptive matching algorithm.The algorithm is preferably user-paired to perform joint particle filter positioning.And it assigns different sound wave ranging frequency bands to different users to avoid mutual interference.Finally,in order to solve the problem of limited client computing power,we designed an indoor positioning system based on WIFI probes.Among them,the router uses WIFI probe technology to proactively detect RSSI information with each terminal.The smart client collects acceleration information and geomagnetic field information.The server performs fingerprint database matching,user pairing,and particle filter operations to achieve separation of information collection,display,and operation.The system stripped the terminal of more work and the positioning efficiency was significantly improved.This article implements system development under Andriod,OpenWrt and Windows systems,and verifies the feasibility of the system and the effectiveness of the algorithm in the actual environment.The algorithm proposed in this paper improves the positioning accuracy when the personnel density is large,and solves the problem of possible mutual interference when multi-user acoustic wave ranging is performed.The positioning system designed in this paper effectively stripped the terminal of a large amount of computation and information collection,providing a feasible system implementation solution for the proposed algorithm.