Unlocking the Potential of Backscattering for Human Activity Pattern Recognition in Smart Environments

The rapid advancement of millimeter-wave (mm Wave) radio-frequency (RF) sensor technology has unlocked new possibilities for tracking human activities indoors. This requires a comprehensive understanding of indoor signal propagation in the presence of moving objects and/or persons. In this context, we present a trajectory-driven non-stationary indoor channel model focusing on backscattering analysis. Expressions for the time-variant (TV) channel transfer function (CTF), the micro-Doppler signatures, and the mean Doppler shift are presented. We perform two distinct experiments in a laboratory environment to extract and analyze the micro-Doppler signatures of a moving person (walking) considering solely the backscattered (double-bounced) signal components. In addition, we also use inertial measurement unit (IMU) sensors integrated in a smartsuit to track the TV trajectories of different human body parts, such as head, chest, and hips. The collected data undergoes rigorous processing to ensure precision and reliability. A good agreement is found between the micro-Doppler signatures computed from the measured RF data and those obtained with the trajectory-driven channel model. The findings also demonstrate the relevance and validity of considering backscattering components to extract the micro-Doppler signature patterns resulting from human motion in a complex indoor environment. This paper offers a comprehensive exploration of the methodology, results, and potential applications, ultimately contributing to the realization of smarter, more efficient, and citizen-centric smart environments.