Novel Microwave Tomography System Using a Phased-Array Antenna

摘要

In this paper, a novel microwave tomography (MWT) system that utilizes a phased-array transmitter antenna is presented. In contrast to contemporary MWT systems, the system presented herein consists of a single transmitter and multiple receivers. The single transmitter is a four-element array antenna; by varying the phase for each element in the array, the field distribution inside the imaging chamber varies, which produces a different set of measurements per phase configuration. Furthermore, the individual elements of the transmitting array antenna and the receiving antennas are of different types; the receiving antenna type is selected to maximize the coupling between the transmitter and the receiver. Due to the system's complexity, the system is modeled using a neural network; the trained network is used as the forward solver for an inversion algorithm based on a Bayesian regularized Levenberg-Marquardt algorithm. The system along with the inversion algorithm is tested using several targets. Furthermore, the performance of the algorithm against various levels of experimental noise is analyzed and evaluated.