Research On Indoor Localization Algorithem Based On Joint Hypothesis Test

With the wide deployment of Wi-Fi network,the Received Signal Strength(RSS)based Wi-Fi indoor localization has been greeted with an embracing of academia and industry.However,most of the existing Wi-Fi indoor localization approaches focus on the goal of increasing localization accuracy,but few of them account for different Access Point(AP)contributions to localization accuracy,and the diversity of Wi-Fi signal distributions as well as the RSS measures involving singular value owing to the complicated indoor environment,which will consequently result in the low robustness of indoor localization system.Based on this,a novel Wi-Fi indoor localization approach based on joint hypothesis test is proposed.The main contents are as follows:Firstly,the normality of Wi-Fi signal distribution at each Reference Point(RP)is tested by the modified Jarque-Bera(JB)test.In offline phase,the skewness coefficient and kurtosis coefficient are calculated by RSS samples at each RP,and a modified JB test statistic is constructed to evaluate the normality of Wi-Fi signal distribution at each RP,accordingly improving the authenticity of the signal distribution at each reference point.Secondly,attribute selection measurement based on Chi-squared Automatic Interaction Detection(CHAID)approach is intended to obtain each AP contribution degree.Specifically,the CHAID approach uses Pearson Chi-square test to describe the location resolution of each AP,so that the AP with higher location resolution is given greater weight.Thirdly,the asymptotic relative efficiency(ARE)of Mann-Whitney U test for T test is theoretically derived.By using the concept of ARE,this thesis illustrates the effectiveness of using Mann-Whitney U test and T test to construct the set of matching RPs under non-normal distribution and normal distribution respectively,which evaluates the performance of the two hypothesis test approaches and reduces the error rate of decision-making judgment,then the hybrid Mann-Whitney U test/T test is proposed.Finally,the joint hypothesis test is used to estimate the location of the moving target.In online phase,according to the result of JB test on Wi-Fi signal distribution at each RP,we apply the hybrid Mann-Whitney U test/T test to construct the set of matching RPs for the moving target,combined with the AP contribution obtained by the CHAID approach,which achieves accurate location estimation of moving targets.The massive experimental results indicate that the proposed indoor localization approach is featured with higher localization accuracy as well as stronger system robustness.