A numerical study of planar elliptical antennas applied to ultrawideband (UWB) imaging of breast tissue

Abstract

The thesis discusses the possibility of using ultrawideband (UWB) radar to detect breast cancer. At the present time, X-ray mammography and ultrasound are the golden standard imaging techniques for detection and evaluation of breast cancer, but they both have their limitations. UWB radar utilizes the difference in dielectric properties between a tumor and the surrounding healthy tissue. By interpreting the reflected signals, it is possible to make a prediction on the localization of a tumor. The primary aim of this thesis is to use ultrawideband planar elliptical antennas in the detection of breast cancer, and investigate the use of the delay-and-sum beamforming technique applied on both the receiver (Rx) and transmitter (Tx).

At first, a description of the antenna and its basic characteristics are given, based on numerical simulations in CST Microwave Studio. This was done in order to get a better understanding of the antenna and its practicability. Then, models consisting of four identical antennas with a layer of fat between them and a tumor inside were constructed. Simulations were performed with the tumor placed on three different locations. The signal-to-clutter ratio is used as a measure of the systems overall performance. Some of the model parameters where changed in order to analyze the systems strengths and weaknesses. These parameters included the noise level, the size of the tumor, the active beamforming process on the transmitter and the difference in dielectric properties between the tumor and the healthy tissue.

General antenna parameters, such as the S11-parameter, radiation patterns, group time delay and phase center were established. Initial indications show that the signal-to-clutter ratios generally increase with the use of the delay-and-sum beamforming technique on the transmitter and receiver, compared to only using it on the receiver. In addition to this, the models using Tx and Rx beamforming seems to have a higher tolerance with respect to changes in the model parameters than the models using only Rx beamforming.