摘要:ANNs for pattern recognition with multi-valued(MV) and multi-thresholded(MT) transfer function as neurons’ nonlinearities have been proposed. The function of multithresholded neurons (MTNs) lay in the fact that it partitioned the input space into several hyperplanar zones by using many thresholds in neurons’ output transfer function, it also marked these different zones with different output values as required. On the basis of these neurons, the advantage of solving parity problem and specific logic problem was investigated. It had been shown that only one neuron was needed to solve N-bit parity problem or to perform arbitrary logic function.Therefore, it was especially suitable for pattern recognition and classification. The combination of MTNs was also investigated. It was found that two MTNs may result in the multiplication of threshold number of combined net.
摘要:A model that combines mass transportation process and surface kinetics is firstly applied to analyze the GexSi1-xCVD process. The temperature, velocity ad concentration distributions in the reactor are calculated by the partial different equations of fluid dynamics. The surface kinetics is used to discuss the deposition of Si and the mass transportation process is used to discuss the deposition of Ge. The theoretical relationship between the initial conditions and the Ge composition x in the solid has been established. The calculated result of the growth rate and the Ge composition in the solid agree with the experiment data. The increase of the growth rate and Ge composition x with the increase of initial GeH4 concentration is explained quantitatively by our model.
关键词:Chemical vapor epitaxy;GexSi1-x;Mechanism of epitaxy
摘要:The transient characteristics of the indirect-coupling photodetector workng in four different regions are reported. When it works in the peak region, the response time is shortened and a new internal current gain is produced. Thus the peak region demonstrates its great value.We have also discussed the experiment result and concluded that the internal current gain is greater than 1 when operatinng in the peak region.
关键词:Photodetector;Indirect-coupling photodetector;response time;Internal current gain
摘要:By means of the effects of one dimension finite well, strain and band-filling, a simple and accurate method to calculate the wavelength of gain peak in compressively strained quantum well lasers is provided. Two key factors which determine the gain peak energy are presented.The influence of Fermi-Dirac distribution function on the wavelength of gain peak is also discussed. Taking commonly-used InGaAs/InGaAsP QW lasers for example, by the use of Lorentzian line shape function, the wavelength of gain peak is calculated under different injection carrier densities.
关键词:Wavelength of gain peak;Feimi-Dirac distribution function;Lorentzian line shape function
摘要:Lossless nonuniform transmission lines are synthesized by the method of characteristics in the time-domain. Given m sample values of the time-domain reflection(TDR) response at m time points, the characteristic impedance profile of a lossless nonuniform line represented by m uniform sections can be obtained. The m uniform sections may be different in length, but have a delay time equal to the relevant sample time interval, The propagation velocity of a signal is not necessarily constant along the transmission line. The TDR response is not split into incident and reflective waves, which makes the synthesis method in this paper more simple and efficient than the traditional synthesis methods based on the scattering parameter theory. One synthesis example is given, showing good synthesis accuracy of the impedance profile.
摘要:This paper aims at the research of the theories and techniques to integrate the intelligent software technologies with the main stream computer technologies, especially with DB technologies, and it aims at providing a multi-paradigm programming language that merges the logic and object-oriented and agent-oriented paradigms into a single whole. We research and implement an expert database system tool suitable for developing different types of expert systems and effectively handling of knowledge and data.
摘要:The criterion which requires each pattern component to possess optimal stability is proposed. Based on the criterion,a constrained perceptron optimization algorithm is developed,which has the following two characteristics: (1) It can store any given training patterns set; (2)Each pattern component possesses optimal stability. Computer simulation results confirm the ad vantages of our optimization learning algorithm over existing ones.
摘要:AlxGa1-xAs/GaAs heterostructure has been grown by two different liquid phase epitaxy(LPE) methods for the fabrication of solar cells. Typical structure characteristics observed under transmission electron microscopy(TEM) and the corresponding device parameters are presented. The results indicate that the P+PNN+ configuration grown by meltback-regrowth method can effectively reduce the defects of the substrate and improve the minority carrier collection by forming a composition-graded region in the window layer. By optimizing the growth conditions and device technology, the best cell performance with conversion efficiency of 18. 78% under AMO,lsun or 23. 17% under AM1. 5, lsun conditions is exhibited.
摘要:It is found in experimment that 1/f noise is closely correlated with negative-bias instability in MOSFETs, which is one of the most important reliability problems for MOS devices.The initial noise spectrum density is proportional to the drift of transconductance of MOSFETs that are subjected to negative bias and high temperature. And, it is shown from the mechanism analysis that both the instability and the noise are induced by oxide charges and traps near Si-SiO2 interface. 1/f noise may therefore be used as an effective tool to predict and analyze negative-bias instability in MOS devices.
摘要:A general small signal transmission line model of superconducting field effect devices is treated analytically. The analysis is based upon an active,distributed transmission line analogy to the kinetic admittance of a superconducting channel region of the superconducting field effect devices. Within the limitation of the gradual channel approximation,a general analysis is presented which is applicable to both superconducting MOSu and PN(or Schottky) junction field effect devices.
关键词:General small signal model;Superconducting field effect devies;superconducting transmission line
摘要:An analytic threshold-voltage model has been established based on the approximate solution of the 2D Poisson’ s equation for the fully-depleted SOI MOSFET. The accuracy of the model is verified by comparison with the simulation results of PISCES and the corresponding experimental data. The model has also such advantages as simple and quick calculation, analytic expression and explicit physical meaning. The establishment of the model is helpful to circuit simulation, study of device physical characteristics and the corresponding process design for the fullydepleted SOI MOSFET.
关键词:Fully-depleted SOI MOSFET;Threshold-voltage;analytic model
摘要:A physical model for the fully depleted short-channel LDD/LDS SOI MOSFET model is developed, which expands our previous thin-film SOI MOSFET model, focusing on channel-length modulation when the SOI MOSFET operates in the saturation region, and the influences of LDD/LDS structure on primary SOI MOSFET I-V characteristics. Our model has a good agreement with experimental results. Moreover, the model describes primary MOSFET model and LDD/LDS structure’ s effects seperately with simple formula and easy parameters’ extraction,which facilitates the implemetaion of this model into circuit simulators such as SPICE.
摘要:A universal structure synthesizer that provides direct analog synthesis of S-do-main transfer functions is presented. To enhance.the circuit performance and production yield,techniques for circuit realization are discussed in detail. Moreover, a piece of advice on layout generation is also given in the paper. The SPICE simulated results of the synthesized Chebyshey filter demonstrate the ability and validity of the proposed approach.
摘要:The key fabricating process of MCT-the triple diffusion is investigated. The process conditions are obtained by process simulator SUPREM-Ⅲ. The measurement results of spreading resistance,bevel and sheet resistance indicate that optimum design and fabrication are fulfilled. The new structure of DOFMCT with turn off capability of 134A/cm2 is designed and fabricated successfully ,
摘要:The noise characteristics of semiconductor laser amplifier are analyzed. The sensitivity formula of receiver with optical preamplifier is derived for the first time. The results show that the input coupling efficiency of optical preamplifier is the main factor that affects the sensitivity of this sort of receiver besides its imner gain, and the optimum judging lecvel is higher than the original receiver’s.
摘要:A Practical macromodel of intcgrated circuit PLL suitable for the general circuit analysis program SPICEE[7] is presented. The macromodel ,which provides a good pin-to-pin representation,simulates very well the output waveforms of LM565 and models satisfactorily the controls of external elements over the performance of LM565. The macromodel can be used in computer-aided-analysis of the circuits containing LM565.
摘要:The first 155Mb/s time-division photonic switching experimental system in China is demonstrated. The key component employed in this system is a new type of semiconductor element-semiconductor optical gate(SOG) . The transmission bit rate of the system is 155Mb/s. The switching among four channels of 34Mb/s users’ data and packet switching of 155Mb/s data are realized. The research work will be a good reference to the research of phtonic data packet swrtching and ATM photonic switching in the future.
摘要:The mixture conductor Cu2Mo6S7.7.Cu2Mo6S8.0 and Cu4Mo6S8.0 of Cu system was obtained under controlled processing condition. It was confirmed to be the single phase similar to the Mo6S8 by X-ray diffraction analysis . and the stoichiometric composition was detected by chemical analytic method. Their total conductivity σ and ionic conductivity σi were measured. The compositon of Cu2Mo6S7.7 and solid electrolyte Rb4Cu16Cl13 is superior to CuxMo6S8-y in electrochemical properties, the total conductivity and ionic conductivity are 75 (Ω· cm)-1 and 7 × 10-2 (Ω·cm)-1 at the composite ratio 2 : 1.
摘要:The current gain of silicon bipolar transistor has been quantitatively modeled at 77K and 300K in consideration of the carrier recombination currents in the emitter-base space charge region and in base,the extra base bandgap narrowing effect by the injection of minority-car-riers at high injection level, the Early effect, the conductivity modulation effects in base and in emitter,the effective base widening effect and the emitter current crowding effect. The obtained results are in agreement with the experimental data. The current gain is mainly determined by the conductivity modulation effect and emitter current crowding effect at 77K at high injection level,but at 300K,it is mainly determined by the effective base widening effect.
摘要:The commonly used methods to measure series resistances in bipolar transistors,including the open-collector method of extraction for emitter series resistance Re, the Ic-Vbe method of measurement of base series resistance Rb and the Forced-Beta method of getting collector series resistance Rc, are analyzed and tested. The results show that the above methods aren’t available to measure series resistances in polysilicon emitter bipolar transistors. A new set of measurement methods of series resistances in bipolar transistors is presented, it overcomes conventional method’ s troubles and is suitable for measurement of series resistances in polysihcon emitter transistors for high speed VLSI.
关键词:Measurement of series resistances in polysilicon emitter transistors;Method of Ning-Tang;Modified Forced-Beta method
摘要:This paper describes a high-frequency and high-power induction heating inverter by using Static induction Transistor(SIT) ,and proposes the technique of choosing and controlling of SIT. The inverter has been successfully used in solid state high frequency reinforcing-bar production line.
摘要:The Pi-sigma network learning algorithm is generalized,and a simple method for computing product of fuzzy numbers is proposed ,so that it is possible to utilize both numerical patterns and expert knowledge represented by fuzzy IF-THEN rules for designing the Pi-sigma network classifiers. Simulation results indicate that the generalized learnig algorithm is correct,and in some cases ,the method based on the Pi-sigma network requires simpler structure and much shorter training time than that based on the Multilayer Perceptron.
关键词:Pi-sigma network;multilayer perceptron;classifier;fuzzy number;α Level set
摘要:Trellis Coded Modulation(TCM) has been widely used because it can obtain 2. 55 ~ 7. 37 dB coding gain without bandwidth expansion or reduction of the effective information rates. Since Viterbi Decoder(VD) plays an important role in the realization of HDTV system,we conduct the research and development on integration of VDs. This paper studies the factors which affect the performance of VDs and their solutions. A completed VLSl circuit design for an 8-state 32QAM VD(30,000 gates in scale) is also reported.
摘要:This paper presents the stability theorem of cellular neural networks(CNN) with a multi-step output function. Using the different steps of the multi-step output function to store different values of pixels,the CNN associative memory of gray-scale patterns is realized.
关键词:Cellular neural networks(CNN);associative memory;Step output function
摘要:An analysis of the causes of bit errors is given for magneto-optical storage systems. For the specificity of different noise intensity in the channel when data of one and zero are read respectively,the bit error characteristics in the channel are modeled. Also, this paper deals with an analysis of the relationship of signal detection method and bit error characteristics in channels. It is shown that the fault freedom in balanced differential channels is better than that in unbalanced differential channels.
摘要:The microwave power coupling and the power absorption by eletron in ECR system are discussed here. It is discovered that the distribution of static magnetic field in ECR resonance cave,the total pressure and the microwave power are the decisive factors for providing high energy electrons. These conditions decide the range of parameters in ECRCVD thin film deposition processes. The practical deposition research of SiN thin film by ECRCVD has been made. The experimental results are consistent with the numerical calculation.
关键词:Electron cyclotron resonance;Microwave power coupling;Power absorption;CVD
摘要:SnO2/CeO2 gas sensing thin films for alcohol were deposited by powder sputtering,whose properties releting to film thickness,temperature and dose are given. For these results above,the technology and structure of plane thin film gas sensors for alcohol were designed. Some properties of these sensors are introduced and the releted problems are discussed.
关键词:Powder sputtering;Thin film;sensor;Gas sensing for alcohol
摘要:The paper accurate by analyses the electeric field of various X-cut Ti: LiNbO3 electro-optic modulators by the finite element method(FEM) ,and calculates the characteristic parameters of them-on/off voltage, modulate bandwidth ,characteristic impedance and effective microwave rsfractive index,etc. We also first consider the thickness of LiNbO3 in calculation of modulator characterisics.
摘要:A self-organizing and self-stabilizing visual pattern learning system,called ART2Cognizer is developed. This system has a good ability of tolerance to shift-in-scale and shift-in-position of the patterns it recognizes.
摘要:The present paper puts forward a new method to realize the temperature compensation,especially,at various levels of input pressure. Relevant circuitry has been worked out,which after being tested at low and high temperatures,proves to be very effective. The derivation of the output/input relationship of the transducer is also given in this paper.