Research On Multi-source And Heterogeneous Fusion Positioning Method

With the rapid development of Internet of Things technology,the ubiquitous heterogeneous network environment provides a rich information fusion space for target positioning.The traditional positioning method based on a single signal source of a homogeneous network has insufficient or inefficient utilization of the information of the environment where the target is located,resulting in the low accuracy and poor robustness of the existing positioning algorithms.In view of the above problems,according to the characteristics of the heterogeneous network environment where the target is located,the wireless local area network(WLAN)is used as backbone network to study a variety of different high-precision fusion positioning methods based on different fusion frameworks.The main content is:1.The inertial navigation aided received signal strength indicator(RSSI)fusion database construction and positioning method are studied.Aiming at the problems of traditional fingerprint positioning in the off-line phase of database acquisition and maintenance workload,and the low reliability of the fingerprint database provided by existing interpolation rapid database building methods,a method of RSSI fusion library building aided by inertial navigation was proposed.The method utilizes multiple data source information from the same WLAN network,including RSSI measurement,inertial sensor measurement,and electronic map information,and can quickly and highly collect the fingerprint database in an actual location environment,and the collected fingerprint database is more robust than the dense fingerprint database provided by other interpolation algorithms.Based on this,a method of dynamic fusion of multiple classifiers and inertial navigation is proposed.Experimental verification in a real environment shows that the dynamic fusion location method is superior to the non-fusion location method and the traditional Kalman filter fusion location method in positioning accuracy and robustness.2.The problem of fusion positioning in heterogeneous network by ultra wide band(UWB)accurate ranging and RSSI measurement in WLAN networks is studied.In view of the fact that the signal is blocked in the actual location environment and the signal propagates mainly in non-line-of-sight,which results in invalidation of time of arrival measurement of most UWB base stations and the inability to meet the actual positioning requirements,a positioning algorithm that combines different measurement information in UWB and WLAN networks is proposed.Through the lognormal model,the algorithm converts RSSI measured in the WLAN into distance information.By constructing an error function optimization model and combining single precise base station ranging,a fast search method is used and achieved accurate positioning using only two WLAN base stations and one UWB measurement.Compared with the existing algorithms,this algorithm has greatly improved the positioning accuracy and timeliness.3.A knowledge-aided RSSI fingerprint and geomagnetic self-adaptive fusion positioning method was studied.In order to fully exploit the heterogeneous information sources in the localization environment,the high-dimensional feature fingerprints with signal level fusion are firstly formed by using the geomagnetic information and RSSI fingerprints calibrated by the inertial sensors,and then the dimensionality of the highdimensional fusion fingerprints is reduced and decorrelationed by the principal component analysis.Secondly,the offline knowledge base is trained,including the training model based on multi-fingerprint function and the global fusion weight matrix.Finally,the off-line knowledge of the above-mentioned fusion weight matrix is used to assist the adaptive fusion positioning on line.The proposed method makes full use of the environmental information of the target,and the results of the experimental test are better than those of a single measurement or a single algorithm.4.The application of dynamic state state detection method based on Channel State Information(CSI)in dynamic fusion location is studied.By introducing environmentsensitive CSI information to detect the presence or absence of personnel in the area to be located,a high-precision regional state detection method is implemented to assist dynamic fusion positioning.The method reduces the matching range of the fingerprint database and greatly reduces the probability of large errors in positioning.Finally,based on Matlab software,a set of positioning platform is built to analyze and present the final positioning results by incorporating the detection information.