Improving the speed and accuracy of indoor localization

摘要: Advances in technology have enabled a large number of computing devices to communicate wirelessly. In addition, radio waves, which are the primary means transmitting data wireless communication, can be used localize 2D and 3D space. As result there has been an increasing applications that rely on availability device location. Many systems developed provide location estimates indoors, where Global Positioning System (GPS) do not work. However, localization indoors faces many challenges. First, system should use as little extra hardware possible, work any with very or no modification, latency small. Also, signals in-doors suffer from environmental effects like reflection, diffraction scattering, making signal characterization respect difficult. Moreover, algorithms require detailed profiling environment, hard deploy. This thesis addresses some aforementioned issues for properties Received Signal Strength (RSS). The advantage these is they reuse existing communication infrastructure, rather than necessitating deployment specialized hardware. Specifically, we improved particular method relies Bayesian Networks (BNs). This requiring small size training data, simultaneously, versions BNs without knowledge locations strength collected. We proposed Markov Chain Monte Carlo (MCMC) evaluated their performance by introducing metric call relative accuracy. reduced identifying MCMC methods improve accuracy solutions returned statistical packages time possible. parallelized process when localizing whose order hundreds. Finally, since transmission heavily affected physical environment investigated impact using multiple antennas various algorithms. showed deploying low-cost at fixed stability indoors.