Positioning technologies for wireless communications

Conventionally, wireless localization schemes work based on the assumption that the sensors/receivers (Rx) are always in Line-of-Sight (LOS) with the transmitter (Tx) to locate Tx. However, this assumption may not be realistic due to multipath, Non-Line-of-Sight (NLOS) propagation and multiple access interference. Under such conditions, the accuracy of localization schemes will be affected. Among these error sources, NLOS is perhaps the most crucial one. In this thesis, we focus on the development and implementation of a robust NLOS mitigation scheme that can improve the accuracy of Time of Arrival (TOA) geo-location in a simulated single moving sensor environment. Our proposed algorithm development can be simplified as 3 structures: Robust adaptive trimming method (basic), Reconstruction of trimmed TOA profile, and Non-parametric (NP) noise density estimator. Note that NLOS errors are modeled as Ɛ-contaminated Gaussian noise with impulsive behaviour and the trimming method is based on a statistical approach to minimize the impulsive noise effect. For further improvement, geometry information between Rx and Tx or polynomial curve fit is suggested to perform reconstruction of trimmed TOA profile. Follow on this; a NP detector which makes minimal a priori assumptions on the noise model, a symmetry density is proposed to integrate with the basic trimming and reconstruction algorithms. Numerical simulations of the above algorithms are done by MATLAB and its results illustrate the promising performance in a mixed LOS and NLOS environment.Bachelor of Engineerin