Study On Indoor Fusion Localization Technology Based On WiFi And Intertial Sensor

In recent years,with the rapid development of mobile communication technology,location information service has been widely used in various aspects of society,and research based on indoor location services has attracted much attention.WiFi positioning has the advantages of low cost and easy implementation.PDR positioning has high positioning accuracy in short time.Therefore,the integration of WiFi and inertial sensors has become the mainstream technology of indoor positioning.This thesis studies the following research on indoor fusion positioning algorithm based on WiFi and inertial sensor:Firstly,aiming at the problem of that the traditional single-core least squares support regression machine positioning algorithm in WiFi positioning has poor fit to RSSI,a WiFi indoor positioned algorithm based on improved multi-core weighted least squares support regression machine(IMK-WLSSVR)is proposed.Firstly,the mapping relationship between RSSI vector and coordinates is built by using IMK-WLSSVR algorithm.Then the improved simulated annealing algorithm is used to optimize the positioning model parameters to improve the positioning performance of the model.Finally,the constructed regression machine positioning model is used to complete the positioning.The simulation results show that the average positioning error of the algorithm is 2.5m,and the positioning accuracy is obviously improved.Secondly,aiming at the problem that the actual positioning accuracy is low in the existing inertial sensor positioning without considering the swaying and pocket mode,a improved pedestrian dead reckoning positioning algorithm based on the recognition of motion pattern and phone gesture is proposed.Firstly,acceleration and gravitational acceleration are used to identify motion mode and phone gesture and selecte the acceleration axis for the step counting,and then the acceleration value peak value is combined with the acceleration threshold value and the time threshold value to detect the step frequency.The gyroscope angular velocity is corrected by the accelerometer error vector,and the heading angle is corrected by using the magnetometer data,compensate the heading angle of the swinging state.The experimental results show that the proposed algorithm achieves accurate step frequency detection and step size estimation in multi-motion mode and phone attitude,which is 23.3% higher than traditional PDR.Aiming at the problem that WiFi wireless signal is vulnerable to environmental influence and PDR method has cumulative error,a WiFi-PDR fusion positioning algorithm based on adaptive EKF is proposed.The adaptive EKF is used to fuse the WiFi positioning with the inertial sensor positioning.The step length and direction angle of PDR module is used to establish the equation of the extended EKF.The WiFi positioning result and the corner position are taken as the EKF observations,and calculate observation covariance to adjust the observed noise covariance matrix in real time.The experimental results show that the results of fusion method is smooth,effectively reducing the rebound phenomenon of WiFi.The positioning error is less than 2m in 92.4% probability.The fusion algorithm is better than single WiFi positioning and PDR positioning.Finally,the software of the positioning terminal is completed and the indoor positioning experiment system is built.The software of the positioning terminal is completed based on the WiFi-PDR fusion positioning algorithm.The positioning experiment system is built by deploying the WiFi AP and carry out positioning experiment test.The experimental results show that the WiFi-PDR fusion algorithm has higher stability in the actual positioning scenario and effectively improves the positioning accuracy.