Indoor Localization with Radio Tomographic Imaging using Channel State Information

Abstract

Abstract This research study aimed to develop a non-learning-based approach for indoor localization using Channel State Information (CSI). The study successfully created a linear model that could perform the task of indoor localization with limited data. The research dataset contained CSI data pertaining to an occupant who was engaged in performing five distinct activities, and each data point in the dataset came with annotations that detailed specific locations and activities. The study also provided valuable insights into how different activities impact CSI in different ways, how amplitude and phase differences produce various outcomes, and how the size of the comparison vector is an important factor in determining the accuracy of CSI data. While the algorithm proposed to locate an individual within a particular area showed promising results, it had some limitations. Overall, the findings of this research study provide valuable contributions and insights that could assist not only in indoor localization but also in activity recognition.