Wireless Localization Based On The Fusion Of Kurtosis Analysis And Channel Model For LOS/NLOS Identification

Positioning and localization is an extensive research area which can be separated in the outdoor positioning system(OPS)and indoor positioning system(IPS).In outdoor environments,for accurate positioning and navigation,we have a standard positioning system termed global positioning system(GPS),but radio signals do not penetrate thick walls and scattered obstacles which result in unpredictable non-line of sight(NLOS)propagation and unreliable estimation of the position.The research community is proposing a wide range of indoor positioning techniques and systems,to fulfill the requirements,but for indoor,there is no standard positioning system like GPS for outdoor.All the positioning developments such as GPS and IPS has entirely transformed the global world but with all advancements,still these systems have restrictions according to their use and availability;which always looked for further expansion.One side GPS is able to provide precise accuracy in outdoor but not feasible in indoor,on the other side available IPS has their own limitations including range and accuracy.In this thesis,we proposed a most popular technology Ultra-wideband(UWB)for indoor positioning due to the high capacity of bandwidth which helps him to resist against multipath and NLOS effects.The UWB system in this research work is used to collect the real-time data and for the identification of the line of sight(LOS)and non-line of sight(NLOS)propagation.Trilateration method is used to find the position of any user and positioning was performed after identifying the LOS and NLOS propagation.UWB system provides the received signal strength(RSS)and first path power level(FPPL)during communication and these parameters are used for the detection algorithm.The measurement results show that how many received signals are LOS or NLOS,furthermore machine learning based approach logistic regression(LR)is also applied for the classification of LOS and NLOS and simulation results illustrates the collected data in binary class.Another primary task was to apply statistical analysis such as Kurtosis and Skewness which demonstrate the tailed and symmetry of the data.The identification was performed on the basis of threshold and after statistical analysis,the gathered data gives us a new threshold that helps to set a range of thresholds with minimum to maximum threshold according to the available indoor environment.