Research on Unsupervised Localization Algorithm of Indoor Target Position under Wi-Fi Signal
The single localization algorithm has the problems of low accuracy and poor stability,and the localiza-tion result of the target trajectory should be obtained directly.In order to solve the above problems,this paper uses mobile sensors(direction sensor,three-axis accelerometer and linear acceleration)to collect data in a 95m×30m ex-perimental environment,and constructs an EPW target trajectory positioning algorithm by integrating the improved Wi-Fi positioning algorithm with PDR optimization algorithm and KEF algorithm.Firstly,the algorithm improves the tra-ditional Wi-Fi positioning algorithm,and uses KNN to optimize the Wi-Fi positioning algorithm based on the analysis results of AP signal loss,which reduces the fluctuation of positioning and improves the accuracy of positioning;then it uses the threshold peak-valley method to detect the step frequency,and optimizes the heading angle estimation through ground conversion,which reduces the cumulative error of the traditional PDR algorithm.Finally,the nonlinear data collected by the improved Wi-Fi positioning algorithm and the PDR optimization positioning algorithm are fitted and fused based on the KEF algorithm,which solves the short board problem of a single algorithm and greatly im-proves the robustness of the positioning system.The simulation results show that when the standard deviation of Gaussian noise is 4.0 dBm,the EPW trajectory localization algorithm has the minimum deviation from the actual traj-ectory,and can still maintain good continuous tracking performance at the corner of the trajectory,with a root mean square error of only 0.97m.The simulation results show that the MEE,RMSE and MAX errors of the EPW algorithm are reduced by 2.82 m,1.76 m and 4.34 m,respectively,compared with the other six algorithms,which indicates that the EPW algorithm has the highest accuracy,stability and robustness.To sum up,the EPW trajectory positioning algo-rithm solves the defects of a single algorithm,improves the accuracy and stability of positioning,and has high simula-tion value.
NETL
NSTL
万方数据
