Principal Component Regression (PCR), Partial Least Square (PLS) and Artificial Neural Networks (ANN) methods are used in the analysis for the near infrared (NIR) spectra of glucose in the whole blood. The calibration model is built up in the spectrum band where there are the glucose has much more spectral absorption than the water, fat, and protein with these methods and the correlation coefficients of the model are showed in this paper. Comparing these results, a suitable method to analyze the glucose NIR spectrum in the whole blood is found.
Near-infrared spectroscopy is a fairly powerful recent but well-established tool that can be used to study and measure biological and chemical concentrations, and the water is an important portion of organism and chemical reagent. Near-infrared (NIR) spectroscopic studies of water have long been a subject of keen interest from various points-of-view. The importance of the NIR spectrum of water stems from the fact that the frequencies and intensities of bands due to water alter with changes in the strength of hydrogen bonds and hydration. The NIR absorbance spectroscopy of water and the single-beam of the water were measured over a temperature range of 10-80°C at increments of 5°C and a spectroscopic range of 1-2.5μm. It had been validated that positions of water absorption bands centered at 6900cm<sup>-1</sup> depend heavily on temperature effects. Temperatures can also influence the bands of the water molecules combinations and overtones. The trend of water absorbance spectroscopy's changing is discussed detailed in this paper. Especially the variety of the band amplitude and the shift tendency in peak position are all presented in this paper. At the same time, the spectroscopies that are collected with different parameter setup of the spectrograph during the temperature change are obtained. Partial least squares (PLS) calibration models were constructed at sixteen separate temperatures.
A novel low-cost 1í+7 plastic optical fiber coupler using cylindrical mixing-rod is proposed. Comparing with conventional tapered mixing-rod plastic optical fiber 1í+7 coupler, the proposed coupler, which uses cylindrical POF as mixing-rod instead of tapered POF in order to save the high demanding POF tapering process, is much lower at cost and only a little deteriorate at crosstalk. The increased crosstalk and its power penalty are theoretically calculated showing that crosstalk lower than 20dB results in insignificant power penalty. The coupler was experimentally tested showing an excess loss of less than 2.5dB and uniformity less than 3.56dB.
A novel Dedicated Multimedia Fiber Industrial Network (FIN) established by consulting the protocols and standards of FDDI-II satisfies the specific requirements of real-time transmission. In this paper, the characteristics of services in FIN are analyzed and the services are classified according to the definition of the Isochronous and Packet ones. Basing on the service partition, the requirements for bandwidth of services are analyzed, and structures of the Cycle Frame and the Token are determined. In the end, the flow charts of transmitting and receiving Cycle Frames in FIN are presented, and the data transaction results of the sub-layers are simulated.
This paper introduces a novel hierarchical model of optical transport network (OTN) and discusses the functional architecture of optical supervisory channel (OSC) with a Q-factor on-line supervision module which can provide a reliable evidence for the transmission quality of operation, administration and maintenance (OA&M) signals in the OSC subsystem of OTN. The quality indication of OSC signals is transferred to the network management system (NMS). The principles of Q-factor detection by a numerical calculation method are presented, and also, an on-line supervisory scheme based on DSP and high-speed A/D chip techniques is set up. The results indicate that the tendency of the calculated Q values matches the actual BER, and it's feasible and accurate to implement Q-factor on-line supervision in OSC of OTN.