A Backscatter Field Model for Near-Field RF Sensing

Backscatter radio has been widely used in remote and wearable sensing for body motion and vital signs. When a moving or shape-changing dielectric is placed in the near-field region of a pair of transmitter (Tx) and receiver (Rx) antennas, the Tx signal propagating to the dielectric will be backscattered and modulated by the motion as a time-varying complex phasor signal at the Rx. In this article, we proposed a semianalytical model based on the electromagnetic field perturbation of the electric dipole (ED) mode of a dielectric sphere. Our proposed model analyzed how the changes in object shape, antenna pair position, and carrier frequency affect the backscattered field. To validate our proposed field model, we designed a torso phantom and performed experiments with an antenna pair at eight different front positions as well as four different sub-1-GHz carrier frequencies. Our model-generated signals had an average of 0.98 cross-correlations with the experimental recording from the six central positions and four frequencies. Furthermore, we added scaling factors to represent the antenna performance for magnitude and phase matching. After performing a least-squares fitting, our scaling model achieved similar fitting errors as the ground-truth measurement variations with average adjusted
$R$
-squared scores of 0.86 and 0.89 for different sensing positions and carrier frequencies.