Optimizing The Indoor Localization Using Channel State Information

The stunning expansion of wireless systems and permeating deployment of wireless techniques in an indoor environment,leads to indoor location based services(LBS)have drenched various traits of contemporary life.In an indoor environment,the most important functionality is to locate the target using wireless devices.Deliberating to the various applications,the developments in indoor localization techniques fall into two broad categories,device-based(DB)and device-free(DF).And in a broader perspective,the device based localization applications requires as a particular device on the objects to achieve the localization of the target.In addition to the device-free localization,the subject does not require carrying a device.Wi-Fi is observed to be an auspicious technique for indoor localization,due to its extensive accessibility and dominant infrastructure in the wireless network.The largest family of Wi-Fi based indoor localization systems depends on received signal strength indicator(RSSI).However,RSSI values are not harmonious due to its rough measurement and high temporal unpredictability.Apart from different from RSSI,the PHY layer power characteristic,channel response can distinguish multipath features,and therefore it retains the likelihood for an accurate and persistent indoor localization.Channel State Information(CSI)has attracted many research efforts in indoor localization,and some prior work has exhibited sub-meter or even decimeter level accuracy.This thesis will utilize CSI to implement high precision indoor localization in complex indoor environment.The main research work mainly falls in two three main parts.At first,it performs CSI calibration on physical layer to remove signal imperfection due to attenuation in channel or hardware.And then a unique MUSIC algorithm is proposed to estimate the Angle-of-Arrival(AOA)and the Signal-Time-of-Flight(STOF).Typically,an indoor environment is complex and changeable,wireless signal produces a multipath effect in the indoor environment.We can obtain many pairs of AOA and STOF of different propagation paths by using MUSIC algorithm.In this Thesis,the k-means clustering algorithm is used to identify the direct path from the multipath components of the signal.After obtaining the AOA and STOF estimation,we treat AOA and STOF as X-value and Y-value in the two-dimensional coordinate system.We can obtain some clusters after using cluster algorithm.At next it defines a likelihood function for each estimated path to be a direct path by analyzing some factors of weighting the direct path and treats the path with the highest likelihood as the direct path.In this thesis,Target discovery using MUSIC(TDUM)system is proposed with optimized angle-ofarrival(AOA),which enhances the potential of commodity Wi-Fi access point to locate a target within tens of centimeter.Numerous experiments are conducted in typical indoor situations with commercial IEEE 802.11 NICs.The experimental results show that the algorithm proposed in this thesis has good performance in estimating AOA and recognizing direct path.In addition,the proposed indoor positioning system has good performance in stability and localization accuracy,and it can be compared with the existing high-precision localization system.