Indoor Personnel Positioning Technology Based On MEMS Sensors And WiFi Integrated

Indoor personnel positioning technology has great application value in many fields,such as pedestrian positioning and navigation,emergency handling,commercial promotion and so on.Because the signal is blocked by the building,the traditional satellite positioning technology is ineffective in the indoor environment.A new positioning technique must be developed for "blindness".In order to obtain the results of indoor positioning,this paper is based on WiFi positioning technology and MEMS sensor positioning technology,and focus on WiFi localization algorithm,MEMS sensor data processing,pedestrian dead reckoning algorithm,pedestrian dead reckoning algorithm with WiFi aided and the fusion localization algorithm of WiFi and dead reckoning.The main research results are as follows:(1)A weighted centroid localization algorithm to eliminate the influence of altitude is proposed considering the influence of AP deployment height on positioning accuracy.Firstly,the RSSI signal are collected and processed,and the distance between the terminal and the AP is obtained according to the indoor signal propagation model.Then,the vertical distance is eliminated by geometric method,and the plane distance between the AP and the terminal is obtained.Finally,the position coordinate is calculated by the weighted centroid localization algorithm to eliminate the influence of altitude.For a multi-storey building,a floor recognition method is proposed,the 2D planar positioning and the floor recognition method is combined and proposed to construct a near-3D indoor location algorithm.The above method is used to identify the floor in the college floor,and the recognition rate can reach 100%.In the large-scale classroom for positioning experiments,the results showed that: compared to the traditional positioning algorithm,the improved algorithm have a certain increase on the accuracy and stability.(2)In view of the high noise characteristics of MEMS sensor,The signal of MIMU are collected and establish ARMA model of MIMU time series and the data signals are processed by Kalman filter.The low pass filter was designed and magnetometer data was processed.And do the calibration and error compensation for magnetometer.(3)The positioning principle of PDR algorithm is analyzed deeply,and the Step counting algorithm,direction angle estimation and step length estimation are studied.A position correction method based on WiFi aided is proposed to eliminate inertial accumulation error.In addition,In current PDR algorithm,the precision of step size estimation is low and the universality is poor,a WiFi aided dynamic adjustment method of frequency-step model is proposed.The parameter of step model is adjusted when pedestrians pass through the correction point two times contnuity.The adjusted model is used for the next phase of PDR.The related experiments were designed in the college building,the results show that the position correction could eliminates the cumulative error of inertial sensors effectively,and after the model parameter was adjusted for the first time,the step size error is reduced by 80%.(4)Based on the idea of integrated navigation,the UKF is used to make loose combination of WiFi positioning system and PDR positioning system.The divergence speed of positioning error of combinatorial system is slower than the single PDR obviously.A new WiFi/PDR fusion location method is proposed based on the WiFi aided PDR algorithm and UKF,and the positioning results are more accurate.