Research On Indoor Localization Algorithm Based On Multi-Source Information Fusion

With the development of the Internet of Things(IoT)technology,indoor positioning services are receiving more and more attention,and people's demand for indoor positioning applications is constantly increasing.However,the complex obstacles in the indoor environment make the indoor positioning signal have a serious multipath fading effect and have complex time-varying characteristics.In order to further improve indoor positioning accuracy and save positioning costs,this paper proposes a multi-source information fusion positioning system in a complex indoor environment.The multi-source information fusion positioning system has two stages.The stage one is the crowd sensing stage,and the stage two is the positioning and navigation stage.In crowd sensing stage,we conduct data collection and fingerprinting establishment for Bluetooth,WiFi,geomagnetism,and computer visual.For the construction of WiFi and geomagnetic fingerprinting,we adopt a compatible walking method for heterogeneous data collection and establish a dynamic fingerprinting,which contains the multi-source sequence information of the indoor feasible route.This method is open to all information sources and simplifies the construction of the database.In the positioning and navigation stage,firstly,the paper proposed the I-Min-Max positioning algorithm based on Bluetooth Low Energy(BLE).This algorithm was improved on the basis of the Min-Max positioning algorithm.I-Min-Max algorithm is a low-complexity and robust positioning method based on the range of Internet of Things(IoT)positioning system.In this paper,we have thoroughly researched the performance of I-Min-Max algorithm under different wireless propagation conditions.The performance is formulated based on the received signal strength(RSS)value available in the IoT device.And we use two different signal propagation models(free space propagation model and lognormal distribution model)to obtain the expected value and variance of the estimation devitation.Then,we proposed an image-based sub-region matching localization algorithm,improved the typical matching algorithm SIFT,improved the feature descriptor,and improved the response time of image matching.In addition,we fuse WiFi and geomagnetic information related to locationinformation to get more precise location estimate.In indoor multi-source information,WiFi,geomagnetism and image information are all available for free,called Wimage.Finally,the experimental results show that,with the assistance of various indoor information,the multi-source information fusion positioning algorithm can effectively improve the positioning accuracy.The root mean square error of the algorithm is only below0.4m,which is very suitable for target positioning in large-scale indoor environment.The multi-source information fusion positioning algorithm proposed is suitable for IoT systems.