Study On Parameter Estimation Method Of Multiple Ultrasonic Echo Signal

Ultrasonic detection is that ultrasonic transducer send ultrasonic pulse to the surface of the measured object, and detects and analyzes the reflected ultrasonic echoes.The purpose is to discover all kinds of useful information about the transmission path and the physical characteristics of the measured object such as its shape, size and so on, It playes a great role in biomedical engineering, industrial non-destructive tests, and defense detection, etc. In order to gain important information from the reflected ultrasonic echoes, it is vital to accurately estimate their parameters.However, due to the existence of noise of transmission medium in ultrasonic testing, the received ultrasonic waves will always have series of changes, and the accuracy of the parameters estimation will degrade in low signal noise ratio. Besides, when it is used to detect multi-layered materials, the reflected ultrasonic waves will overlap as the increase of the material layers and the decrease of the material thickness. Thus, it is difficult to gain the information about each reflected wave from the overlapping ultrasonic echoes. In that case, it is of great value for the ultrasonic testing method to be able to separate those overlapping echoes.This thesis is focused on the separation and detection of the overlapping reflected ultrasonic waves in low signal noise ratio condition.The work and the innovative achievements of this thesis are as follows:1.By combining the wavelet transform method with the ABC (Artificial Bee Colony) and appling them into the estimation of the parameter of reflected ultrasonic waves. We firstly use the wavelet transform method to process the reflected ultrasonic waves, and then utilizes the ABC to estimate those parameters.The accuracy of this hybrid optimization method is much high. Compared with other algorithms, the method also works well in the estimation of each wave from the overlapping reflected ultrasonic waves in low signal noise ratio condition.the numerical simulation prove that the mean square error of this hybrid optimization method is close to the CRLB (Cramer-Rao Lower Bound).2. After estimate the parameters of the reflected ultrasonic echoes, the AIC information criterion is applied to estimate the number of the overlapping reflected ultrasonic waves by using their estimated arrival time and amplitude correction. The results demonstrate that this method can achieve the number of overlapping reflected ultrasonic waves with higher accuracy. And with the increase of low signal noise ratio, the accuracy upgrades. But with the increase of the number of reflected waves, the accuracy of the number degrades.3. It is still challenging to accurately detect severed overlapping reflected ultrasonic echoes. This thesis puts forward the use of linear modeling method and effectively separates the overlapping reflected ultrasonic waves in thin layer materials.4. The pulse echo in water tank for different samples.The hybrid optimization method is used to estimate the parameter and the number of the reflected echoes. The experiment verifies the validity and utility of the hybrid optimization method, which will have certain practical value in the field of biomedical engineering.