Pan Tianhong

Statistical key variable analysis and model-based control for improvement performance in a deep reactive ion etching process

This paper proposes to develop a data-driven vias depth estimator of the deep reactive ion etching process based on statistical identification of key variables. Several feature extraction algorithms are presented to reduce the high-dimensional data and effectively undertake the subsequent virtual metrology (VM) model building process. With the available on-line VM model, the model-based controller is hence readily applicable to improve the quality of a vias depth. Real operational data taken from a industrial manufacturing process are used to verify the effectiveness of the proposed method. The results demonstrate that the proposed method can decrease the MSE from 2.2 ? 10?2 to 9 ? 10?4 and has great potential in improving the existing DRIE process.

The application of neural networks in smell analyzing system of grain quality

Moldy grain have mildew which are harmful to people and animals. To provide a simple and objective solution to identify whether grain are moldy, a smell analyzing system of grain quality is developed. The system consists of gas sensor array, gas pipe system, signal adjusting circuit, data collecting system and pattern recognition system. The software is developed on the basis of Visual Basic 6.0 and Matlab 6.5. In order to improve veracity of identification, three eigenvalues in the collecting data are used, one is the maximal response point, the other two are near the maximal response point. The sample eigenvalues are trained with the triplex-optimize BP NN. By testing, both training samples and test samples can be identified correctly, which indicates that the application of neural networks is valuable in the smell analyzing system of grain quality.