| With the rapid popularization of wireless network and mobile terminals,the application technology of location-based service continues to mature,which becomes indispensable in daily life.The WiFi indoor positioning technique based on the fingerprint has many advantages such as high positioning accuracy,low positioning cost,and easy deployment,which attracts more and more attentions.In addition,with the rapid development of microelectronic technique and micro electro mechanical systems(MEMS),the mobile terminal integrates massive high-precision sensors,which makes the inertial sensor-based indoor positioning possible.Compared to other positioning techniques the inertial sensor-based indoor positioning technique has strong resistance to the external interference.However,as the error of gyroscope and the direction sensor will accumulate over time,it will result in the decline of positioning accuracy.According to the characteristics of WiFi and inertial sensor for the indoor positioning,this thesis improves the accuracy of indoor positioning by combat their shortcomings.The main contributions are following:(1)Aiming at the problem of the hard work to build the position fingerprint database for the WiFi-based indoor positioning,this thesis proposed a position fingerprint construction algorithm based on the kriging interpolation which is the improvement of Tyson polygons.By utilizing the feature of equal distance from the vertex of Thiessen polygon to the discrete points,the compuation complexity to obtain the variation function is reduced.To eliminate the influence of randomness selection of the variation function model,the improved least square method is used to fit the function expression,and the position fingerprint of the point to be interpolated is obtained.Simulation results show that,the error is 1.345 m when the WKNN algorithm positioning is used with the unfilled database,while the error is 1.058 m when the WKNN algorithm positioning is used with thefilled database.Therefore,to ensure the positioning accuracy,compared with the point-by-point method,the workload of this method is reduced by 45%.(2)For the pedestrian gait detection in indoor positioning PDR algorithm based on inertial sensors,some improvements have been made for the step size estimation and heading angle estimation.To detect the gait,a dynamic extremum difference threshold has been proposed,which utilizes the extremum difference to achieve the dynamic estimation of gait,and cancels the error accumulation duration the estimation of the heading angle.Experimental results show that,theaccuracy of gait detection is increased by about 10%,and the accumulation of positioning error has been effectively suppressed,accuracy of PDR positioning algorithm has been improved.(3)Based on the extended Kalman filter,the fusion of WiFi fingerprint and PDR was proposed for the indoor positioning,in which the state equation and observation equation of the positioning system are established respectively.Simulation results show that the average error is45.3 % lower than that of the PDR positioning algorithm,and 57.9% lower than that of the WiFi positioning algorithm.It verifies such solution has advantages in accuracy and stability. |
