Present Research

	Doppler ultrasound is a non-invasive technique for measuring the blood flow inside the human body. It is significant to make research on new method and technique of Doppler ultrasound for the benefits of the patients and the convenience of the physicians. My research is made up of two parts of work on the whole:
	The simulation of Doppler ultrasound signal Simulations have some advantages over experiments in the study of signal analysis techniques. A simulation on a computer can usually be done more quickly and less expensively and the simulated signal is free from the noise caused by the ultrasound instrument or the human body itself. Furthermore, if the simulation is based on physical model, many instrument parameters can be modeled and device design can be altered if necessary without the need to construct a prototype, so our work focus on the physical model. At first, the model is limited to the steady blood flow. Then it is extended to the laminar, pulsatile, nonuniform flow. Now, the signal from a vortex can be simulated. Our final purpose is the simulation of the blood flow through blood vessel with stenoses, because it is very important and familiar in clinical medicine. It still needs further efforts. This part of work has been presented as a series of papers.
	The processing of Doppler ultrasound signal It includes the modeling, the feature extraction and the classification of Doppler ultrasound signal. Three models are introduced so far: transmission line model of the vessel, zero-pole model and transfer function model of the signal. Analysis of the zero-pole model for the umbilical blood flow reveals that some parameters of the model are sensitive to some gravidity diseases and the sensitivity is more obvious than the conventional ones. The parameters can be used as clinical criterion. Some results of the other two models are also been acquired. Considering the nonlinear characteristic of Doppler ultrasound signal, I presume that the result will be more persuasive if I add nonlinear consideration into the models. For example, substitute the linear zero-pole model with a bilinear model, the simplest nonlinear model. This will be my next attempt. Nonlinear characteristic of the Doppler signal has been validated by my predecessor. Fractal dimension, together with some parameters from other models, will be tested as the classification criterion of some diseases later.