Identification of Low Observable Targets in Through Wall Imaging(TWI) Using Parallel Analytical Compressive Sensing (PACS) method

Through Wall Imaging (TWI) can be implemented using compressive sensing (CS) with fewer measurements, and a real model scenario of TWI can be realized using the finite difference time domain (FDTD) method. We propose parallel analytical compressive sensing (PACS) method to improve dielectric objects detection through the wall. We also present an optimized convolutional perfectly matched layers (CPML) using artificial bee colony (ABC) algorithm to reduce low frequency harmonics reflection of the ultra-wide band (UWB) transmitted signal from CPML layers. The PACS method calculates the measurement matrix based on the local maximums of the received signals. This process involves converting the norm-1 cost function into norm-2, providing a closed form formula to solve the CS problem. Coherence condition is satisfied by singular value decomposition (SVD) of the measurement matrix which in turn reduces the hardware complexity of the receiver. Closed form formula is executed separately for each target using a novel algorithm and their answers are added up. The proposed method achieves a compression ratio of 15% and is compared with the CFAR-BP algorithm for different SNRs which demonstrates its performance accuracy.