Research On Mobile Indoor Positioning Technology Based On Multi-source Data Fusion

As the concept of location-based service(LBS)is proposed,the demand for indoor location service in the location service system is increasing.In recent years,a variety of indoor positioning technologies and applications are proposed in the field of indoor positioning,including UWB-based positioning technology,Bluetooth-based positioning technology,WiFibased positioning technology and indoor positioning technology related to inertial navigation,etc.However,each indoor positioning technology has its own advantages,and single positioning technology is difficult to have universality.And the combination of multiple positioning technologies is the main solution of indoor positioning.Therefore,this paper studies the WiFi fingerprint location technology and the location technology based on the inertial measurement unit.Then the experiments of the positioning methods are carried out and analized.Morevoer,the indoor location technology based on multi-source information fusion is proposed.The main research contents are as follows:1.In view of the difficulty of off-line WiFi intensity fingerprint information collection in WiFi fingerprint location technology,this paper introduces the traditional grid method to collect and construct WiFi fingerprint database,and puts forward semi-simulation construction method,target recognition construction method based on binocular vision,mobile car collection method,then the Kriging interpolation method is applied to generate the final WiFi location fingerprint database,which reduces the work of off-line collection of WiFi fingerprinting information.2.For the WiFi online location algorithm,this paper proposes a fingerprint matching algorithm based on random forest and improved KNN,which reduces the computation in the online location stage and eliminates the influence of other indoor WiFi fingerprint information.The experimental results show that the algorithm is more accurate than other traditional WiFi fingerprint matching algorithms.3.For the problem of processing the output information of the built-in inertial measurement unit in the intelligent mobile terminal,this paper introduces the PDR algorithm used in the relative positioning of the mobile terminal,and the gait detection algorithm is improved in the PDR algorithm.In addition,the method of angle correction is put forward.Moreover,the acceleration information is processed by using the sliding window.And the relevant features are extracted for the pedestrian motion pattern recognition.The experimental results show that the recognition accuracy is up to 91.8%.And the results of pedestrian motion pattern recognition can be used to modify the location results.4.To solve the problem of multi-source information fusion,a particle swarm optimization algorithm based on multi-source information fusion is proposed,which combines WiFi location information,PDR location information,Bluetooth strength information and map constraint information.And the closed path experiment in indoor environment is designed.The combined location algorithm,single WiFi fingerprint location algorithm and PDR location algorithm are compared and verified,and the location errors are analyzed.The experimental results show that the combined location algorithm proposed in this paper is more accurate,whose average positioning error is 62.1% lower than that of the single WiFi fingerprint location algorithm,and 88.5% lower than that of PDR location algorithm.The direction update error in PDR relative positioning can be reduced effectively by the map constraint information,and the accumulated error in the process of INS positioning can be reduced by WiFi fingerprint positioning results and Bluetooth strength information.Therefore,the intergrated positioning error is smaller.