Research On Indoor Fingerprint Localization Method Based On Multi Source Information Fusion

With the development of human life and production,the demand for indoor positioning has become more intense.In greenhouse,indoor positioning technology can be used to locate operators and guide them remotely,which can improve work efficiency.In agricultural garden,it can be the planning path of agricultural robot.In the agricultural tour,it can provide tourists with real sightseeing navigation.In the comprehensive market,it can provide navigation services for customers and so on.Based on the above analysis,this thesis presents a multi-source information fusion indoor fingerprint positioning method,through the integration of intelligent devices,multi-sensor(geomagnetism sensors,barometric sensors,acceleration sensors)data to achieve high-precision positioning.Among them,the data collected by the geomagnetic sensor is used to match the position where the positioning target is located.The acceleration sensor and the direction sensor are used to predict the position that the positioning target can reach at the next moment,and the barometric sensor is used to judge the floor change.Specific positioning method can be divided into three steps: First,collecting the geomagnetic data of the indoor positioning area,and constructing the geomagnetic fingerprint database of the positioning area.Second,the dynamic time warping algorithm(DTW)algorithm is used to match the real-time acquired geomagnetic data with the geomagnetic fingerprint database to obtain the position of the target and then the acceleration and direction sensors are used to predict the position that the target can reach at the next moment,and based on the data collected by the barometric sensor to achieve the floor to determine the initial positioning.Third,K-means clustering algorithm is used to optimize the positioning results to improve the positioning accuracy.Compared with the traditional indoor fingerprint localization method,the positioning method proposed in this thesis takes geomagnetic data as the positioning feature,and the geomagnetic fingerprint database is very stable because the geomagnetic data in indoor environment are mainly affected by large metal(such as elevator,etc.).Moreover,the method proposed in this thesis fuses accelerometer and barometric sensor,the acceleration sensor and direction sensor can accurately predict the location of the next possible target location.Therefore,a small range of geomagnetic fingerprint matching based on the predicted results can reduce the amount of calculation and improve the positioning accuracy of the intelligent terminal.The barometric sensor can collect pressure data,with the change of target height,the barometric sensor can accurately identify the floor,so the fusion pressure sensor can realize the multi floor judgment.Therefore,the multi-sensor fusion indoor positioning method can achieve high accuracy,low computational complexity,reliable positioning of multiple floors.Finally,in order to verify the accuracy of the localization algorithm,the localization method proposed in this thesis is deployed into the intelligent device.Experimental results show that the accuracy of using DTW algorithm is 96%.After Kmeans clustering method is used to eliminate the error,the positioning accuracy was improved by 0.6m.The average positioning error of the system is about 1.4m,so it can meet the positioning accuracy requirements of most indoor scenes.