Ultra-Wideband Time Domain Measurement Systems for Microwave Tomography
Licentiate thesis, 2010
As a potential imaging modality for biomedical applications, microwave tomography has attracted a lot of interests in recent years. With microwave tomography, microwave signals are transmitted into the biological tissues and the scattered fields are recorded. By making use of the received information, the dielectric property profiles of the tissues are quantitatively reconstructed. Based on the difference in dielectric properties, tissues are differentiated and
consequentially can be imaged.
It has been shown that the reconstructions from wideband data present very good quality with high spatial resolution and high accuracy. In our previous work, the wideband data were obtained in frequency domain with a vector network analyzer, which was time consuming and expensive. An
ultra-wideband time domain measurement system has the advantage of fast measurement of wideband data, which is preferable for the clinical applications.
A time domain measurement system is either based on equivalent time sampling or real time sampling. In this thesis, the system architectures and work principles of ultra-wideband time domain measurement systems based
on these two different sampling strategies are described.
In comparison with a frequency domain measurement system, a time domain system, especially that with real time sampling, has lower measurement accuracy due to the poor signal to noise ratio. In this thesis, the accuracy analysis of time domain measurement systems are carried out, which includes
the identification of the error sources, the modeling of different types of errors and the terminologies of interest for evaluating the measurementaccuracy. The effects of measurement error on the image reconstruction is
also investigated by taking a 2-D breast model as an example.
Last, an ultra-wideband time domain experimental system for microwave tomography is presented. Its measurement repeatability is evaluated and phantom reconstructions made from the time domain measurements are compared with those obtained by using a wideband frequency domain system
Real Time Sampling
Frequency Domain Measurement