Human Identification Using Near-Field Bi-Static Radar at Low Frequencies.

Near-field scattering of human targets in the view of a bi-static, radar-like sensor operating in the lower radiofrequencies is used as an alternative to traditional biometric identification systems. These radiofrequency-based human sensor systems have emerged as a promising solution to address privacy concerns, particularly those associated with audio and visual data that extract sensitive personally identifiable information. In this paper, we propose a novel method for privacy-preserving human identification using bi-static radar-like sensors. Unlike conventional radar systems that rely on echoes and reflections in the far field, our approach is based on the transmission of signals through and around users as they pass through a transmitter and receiver. Instead of the more commonly used linear or segmented swept frequencies, this work utilizes discrete swept frequencies to transmit and receive radiofrequency signals. We have examined the performance of seven machine learning models in terms of accuracy and processing time and found that the Extra Trees ensemble model produced the best results, with an accuracy rate of 94.25% for a sample size of 31 individuals using an Intel(R) Core(TM) i5-10300H CPU @ 2.50 GHz processor. [ABSTRACT FROM AUTHOR]