Modeling of Near-Field Self-Injection-Locked Radar for Wrist-Pulse Sensing: Small-Signal Analysis and Validation

Abstract

Near-field wrist-pulse sensing using self-injection-locked (SIL) radar enables accurate detection of small physiological motion and is, therefore, attractive for vital-sign detection. However, existing near-field SIL wrist sensors are still designed largely empirically. Antenna optimization often relies on qualitative field inspection or conventional port metrics, and system-level analyses commonly stop at confirming that dielectric perturbations induce frequency/phase modulation. As a result, a practical and measurement-oriented link from antenna near-field characteristics to the final demodulated baseband output level remains insufficient for design guidance and sensitivity budgeting. This work develops a design-oriented small-signal framework for near-field SIL wrist-pulse sensing. The framework combines the antenna-side perturbation term, injection-locked oscillator response, slope-detection conversion, detector conversion, and baseband gain into a calibrated sensitivity-decomposition chain. In addition, a target-region-oriented near-field electric-field figure of merit (NEFOM), combining effective field coverage and phase coherence, is introduced as a design-stage tool for comparing sensing antennas in near-field wrist-sensing scenarios. The proposed framework is evaluated using a benchmark antenna set and further validated on a fixed SIL radar platform using two 5.8-GHz complementary split-ring resonator (CSRR) sensing antennas with similar matching conditions but different near-field metric levels. The measured results are consistent with the derived small-signal framework.