Signal propagating through free space in wireless communication is subject to additive noise by line-of-sight and non-line-of-sight objects in the propagation medium. This leads to a lot of interference and scattering due to multipath effects. This research work aims to identify such contributors in the propagation channel and characterize them based on their signal scattering property. A data-driven modelling approach is used in place of the traditional math-based modelling. K-means clustering along with other data interpretation methods were used to identify the scatterers. The scatterers are either characterized as absorbing or reflecting type based on the way the signal is affected. Five independent datasets using the C-band frequency were collected under laboratory conditions and used for the study. The ideal dataset from the manufacturer was used as the benchmark. The results identified the scatterers from the experimental dataset and enabled the estimation of their dimensions and material composition in laboratory conditions.
H. I. Surej, Karthic, S., Vigneshwara, G., Jeyashri, T., Dr. T. Rajagopalan, Gandhiraj, R., Kumar, K. A. Pradeep, Binoy, B. N., Sundaram, G. A. Shanmug, and Ram, D. S. Harish, “Evidence of Scatter in C-band Spatio-temporal Signals using Machine Learning Models”, 2020 International Conference on Communication and Signal Processing (ICCSP). 2020.