Human Detection Based on Learning and Classification of Radio Scattering Parameters and Para-Hermitian Eigenvalue Decomposition

Frank E. Ebong, Nicola Novello, Andrea M. Tonello

Abstract: In this article, we introduce a novel machine learning algorithm termed the S-parameters-based Human Detection Algorithm (S-PBHD). Designed to operate on radio frequency datasets transformed into an S-parameter multi-port channel matrix, our algorithm aims to detect human targets within indoor environments using distributed antennas. Central to the efficacy of this algorithm lies its preprocessing phase, which entails the extraction of critical components from the scattered radio field. This is achieved through the Para-Hermitian Eigenvalue Decomposition method, which effectively disentangles the complexities inherent in the S-parameter multi-port channel matrix.