Advancing Indoor Radar Sensing: Non-Stationary Channel Modelling, Cross-Section Estimation, and Validation Through Arm Movement Experiments

In this research, we introduce a novel trajectory-driven channel model for non-stationary indoor millimeter wave (mmWave) communication environments. The model incorporates optimized time-variant (TV) path gains, particularly focusing on scenarios with multiple moving point scatterers. Determining optimal TV path gain values for each moving point scatterer requires estimating the respective radar cross-sections (RCSs). To address this challenge, we formulate an optimization problem based on the $L_{p}$-norm, utilizing the TV mean Doppler shift as a key parameter. We then estimate RCSs by fitting the TV mean Doppler shift of the channel model as closely as possible to that computed from real radar measurements. The efficacy of this approach is validated through experiments in a controlled laboratory setting using a 24 GHz radar kit. Two distinct experiments with right arm movements are examined, allowing us to estimate the RCSs of various arm segments; hand, upper arm, forearm, and shoulder. Our numerical results demonstrate the accuracy and effectiveness of the proposed methodology in emulating real radio channel conditions.