Spatiotemporal Parallel Remote Sensing of Multiple Adjacent Concealed Human Targets

Remote sensing of concealed targets plays a pivotal role in critical fields such as earthquake rescue and military operations. The resolution of targets is constrained by the radar baseline length, which significantly degrades the accuracy of detecting multiple adjacent concealed targets. To address this challenge, we propose an accurate and robust spatiotemporal parallel processing algorithm that overcomes the limitation of spatial resolution. For ultra-wideband (UWB) multiple-input and multiple-output (MIMO) radar, the relationship between echo signals and the spatiotemporal characteristics of targets is derived in detail. Subsequently, the optimization of features in weak human echo signals is formulated as a least absolute shrinkage and selection operator (LASSO) problem. In our scheme, the problem is solved by integrating forward and backward spatiotemporal sensing factors with alternating direction method of multipliers (ADMMs). We conducted comprehensive simulations and experiments, demonstrating that the proposed algorithm can simultaneously extract the positions and respiratory rates of multiple human targets. At a range of 4 m, an azimuthal resolution of 40 cm for adjacent targets is achieved, effectively surpassing the limitations of array spatial resolution. Moreover, both simulations and experiments confirm that the algorithm exhibits strong adaptability to low signal-to-noise ratio (SNR) environments, with SNR improvements exceeding 10 dB.