Implementation of N-Step-Ahead neurocontrol on a 3-Axes heavy duty hydraulic manipulator

An N-step-ahead control method called Chain-Back-Propagation (CBP) is presented. The method is a systematic integration and extension of the truck-backer-upper example. It uses a pair of neural networks, various global and chain-local penalty functions and convergence control limitations. The CBP method is implemented on a real-world hydraulic Cartesian 3-DOF heavy duty manipulator. The results, compared to conventional and extended PD control, are quite satisfactory, indicating improved accuracy and faster response. © 1998 Elsevier Science Ltd. All rights reserved.     

基于神经网络的单步和多步电弧炉模型的研究与应用

介绍了电弧炉计算机智能控制系统的特点与方法，并着重讨论了神经网络预估模型的设计，包括遗传算法在建立神经网络模型的应用和神经网络模型实时学习的要求，基于此基础对于多步预估进行了讨论和研究。通过实际应用证实该系统具有效率高，电能消耗低，电极消耗和耐火材料消耗少等一系列优点。     

A neural network approach to the traffic control problem in reverse baseline networks

A neural network for the traffic control problem applied to reverse baseline networks has been proposed in this paper. This problem has been first represented by an energy function. A neural network is applied for maximizing the energy of the function under the constraints of the reverse baseline network. The number of iteration steps in our neural network is limited by a performed upper bound O(n), wheren is the size of ann ×n network. The throughputs of our neural network have been shown by the empirical results to be better than the conventional algorithm (modified Bipartite Matching Algorithm) when the packet densities rise higher than 50%.     

Time-varying spectral estimation using AR models with variableforgetting factors

A method of estimating time-varying spectra of nonstationary signals using recursive least squares (RLS) with variable forgetting factors (VFFs) is described. The VFF is adapted to a nonstationary signal by an extended prediction error criterion which accounts for the nonstationarity of the signal. This method has better adaptability than the conventional algorithm with high fixed forgetting factor (FFF) in the nonstationary situation, and has lower variance than the conventional one with low FFF in the stationary situation. The extra computation time for the forgetting adaptation is almost negligible     

A numerical analysis of the d.c. performance of small geometry lateral transistors

A numerical analysis technique is employed to determine the performance of small lateral geometry transistors in silicon integrated circuits. Results presented include the base transport factor and common-emitter gain of lateral transistors operating at current levels below the onset of base-region high injection. An n⁺ buried layer is introduced and shown to give an improvement in base transport factor but not necessarily in common-emitter current gain. Computer-generated plots indicating areas of high recombination and the distribution of current with the base of the transistor illustrate this effect. The predicted performance of a transistor with an n⁺ buried layer is compared with data from production devices.The program uses a finite-difference method to solve the minority carrier transport and continuity equations within a two-dimensional rectangular mesh. The dependence of carrier mobility on electric field and impurity concentration magnitude is incorporated. The program uses over-relaxation, the optimum factor being determined by Carré's method.     

A new dielectric isolation method using porous silicon

A new dielectric isolation technology is proposed. In the new structure, single crystalline Si islands are separated from the silicon substrate by oxidized porous silicon. It is based on the following characteristics of the porous silicon oxide formation: (1) p-type Si is more easily changed to porous silicon than n-type Si; (2) porous silicon is formed along the anodic reaction current flow line; (3) the change in volume of porous silicon after oxidation is relatively small; (4) thick porous silicon films (10 μm) can be obtained easily. In this method, a p-type isolated layer is obtained by proton implantation used for an n-type layer formation. Lateral p-n junctions fabricated in such isolated silicon layers show lower leakage current than those reported in SOS technology.     

Cloud security situation prediction method based on grey wolfoptimization and BP neural network

Aiming at the accuracy and error correction of cloud security situation prediction, a cloud security situation prediction method based on grey wolf optimization (GWO) and back propagation (BP) neural network is proposed.Firstly, the adaptive disturbance convergence factor is used to improve the GWO algorithm, so as to improve theconvergence speed and accuracy of the algorithm. The Chebyshev chaotic mapping is introduced into the positionupdate formula of GWO algorithm, which is used to select the features of the cloud security situation prediction dataand optimize the parameters of the BP neural network prediction model to minimize the prediction output error.Then, the initial weights and thresholds of BP neural network are modified by the improved GWO algorithm toincrease the learning efficiency and accuracy of BP neural network. Finally, the real data sets of Tencent cloudplatform are predicted. The simulation results show that the proposed method has lower mean square error (MSE)and mean absolute error (MAE) compared with BP neural network, BP neural network based on genetic algorithm(GA-BP), BP neural network based on particle swarm optimization (PSO-BP) and BP neural network based onGWO algorithm (GWO-BP). The proposed method has better stability, robustness and prediction accuracy.     

Modeling and Optimization of Thermoelements by a Combined Analytical and Numerical Method

A combined analytical and numerical process has been developed to model and optimize thermoelements. In this way, the performance of commercial n- and p-type thermoelectric materials can be optimized to deliver the maximum output power and conversion efficiency. The validity of the method is demonstrated using a silicon germanium unicouple.     

神经网络在电网负荷预测上的应用研究

研究了人工神经网络方法,提出使用BP神经网络建立短期负荷预测综合模型.为加快神经网络的学习速度,改进BP算法,引入平滑系数和遗忘系数,并取得良好的预测效果.     

Striations als ursache von knickstellen in pn-//übergängen von leistungsbauelementen

On chemically etched cross-sections of high-voltage power thyristors both p-n junctions adjoining the slightly doped middle zone revealed numerous kinks which appeared more frequently at the rim of the cross-section than in the middle section. These variations of the p-n junction could be related to striations which in this case are caused by microscopic variations of the phosphorus concentration of the used starting silicon. It is shown that striations intersect the p-n junction at such kinks and that in this case striations are locations of abrupt resistivity variations of the silicon. The deviation of the p-n junction at such kinks caused by striations can amount to 10 μm. This corresponds to a variation of the resistivity (//?_max/?inmin) by a factor of 3. Recent results from spreading resistance measurements of the ?-variations in striated silicon are compatible therewith.     

Characteristics of metal-semiconductor contacts fabricated by the electroless deposition method

Metal-semiconductor contacts have been fabricated by electroless deposition of Cu on chemically cleaned n-type silicon and their characteristics studied. The values of barrier height and the ideality factor are found to be comparable to those of vacuum evaporated contacts. A non-linearity in the 1/C² vs V plot has been observed and the same has been satisfactorily explained by taking surface state capacitance into consideration.     

基于小波神经网络非线性预测方法的研究

针对非线性预测问题,提出了小波神经网络算法．该算法采用权重贡献率分析法和关键神经节点法分析权重,精进模型,利用具有优良渐进性的递推预报误差法训练小波的尺度因子和平移因子,并提出了一种网络的改进算法．通过对导航设备的仿真预测,该算法优于同等规模的BP神经网络,其收敛速度快,预测精度高．     

A p-channel MESFET on silicon using an erbium gate

A p-channel MESFET (metal semiconductor field-effect transistor) has been fabricated using erbium as gate material and iridium as source and drain contacts. The results show that it is possible to achieve p-type devices with characteristics comparable to n-type devices. As substrate silicon on sapphire (SOS) was used since it gives a well-defined channel thickness. The thickness of the silicon was 0.6 μm; after processing this was reduced to about 0.5 μm, which was, thus, the ultimate channel thickness.     

A computer-aided simulation model for the I–V characteristic of M-n-p silicon Schottky-barrier diodes produced by use of low-energy arsenic-ion implantation

Current-transport properties of Al-n-p silicon Schottky-barrier diodes have been studied both experimentally and theoretically. An analytical model for the I-V characteristic of a metal-n-p Schottky barrier diode has been developed by using an interfacial layer-thermionic-diffusion model. Assuming a Gaussian distribution for the implanted profile, the barrier-height enhancement and ideality factor have been derived analytically. Using low energy (25 KeV) arsenic implantation with the dose ranged form 8 × 10¹⁰/cm² to 10¹²/cm², Al-n-p silicon Schottky barrier diodes have been fabricated and characterized. Comparisons between the experimental measurements and the results of computer simulations have been performed and satisfactory agreements between these comparisons have been obtained. The reverse I–V characteristics of the fabricated Al-n-p silicon Schottky barrier diodes can also be well simulated by the developed model.     

硅基毫米波共面波导可缩放神经网络模型

随着微电子工艺技术的发展,硅基CMOS器件的截止频率已经达到毫米波频段,使硅基微波单片集成电路实现成为可能。因此,建立硅基毫米波频段共面波导结构模型使准确设计硅基微波单片集成电路成为必要。文章提出了一种基于神经网络技术的共面波导结构(CPW)毫米波可缩放模型,采用3层神经网络结构,根据共面波导的测试结果,用神经网络来学习其物理变量和测试的相应S参数空间映射关系。仿真与测试结果比较表明:基于神经网络方法建立的毫米波共面波导可缩放模型对不同几何参数CPW能够快速和准确地给出对应的CPW的S参数结果。     

基于外界影响及时序因素的PM2.5预测研究

随着雾霾问题逐渐加重,对其主成分之一PM2.5的预测已成为广泛关注的问题.PM2.5日浓度变化受多种因素影响,且具有非线性、时变性的特征,难以被准确预测.针对此问题,文中提出一种基于外界影响及时序因素的PM2.5日浓度预测方法.该方法分离出PM2.5日浓度的外界主要影响因素与时间因素,建立了基于外界主要影响因素的BP神...     

Electric Current Dependence of a Self-Cooling Device Consisting of Silicon Wafers Connected to a Power MOSFET

A self-cooling device has been developed by combining a commercial n-channel power metal–oxide–semiconductor field-effect transistor (MOSFET) and single-crystalline Sb-doped n-type or B-doped p-type silicon wafers in order to improve the heat removal or cooling quantitatively. The electric current dependence of the temperature distribution in the self-cooling device and the voltage between the source and drain electrodes have been measured to estimate the Peltier heat flux. We found that the average temperature is decreased for a power MOSFET in which an electric current of 50 A flows. In particular, the average temperature of the power MOSFET was decreased by 2.7°C with the n-type Si wafer and by 3.5°C with the p-type Si wafer, although an electric current of 40 A makes little difference. This certainly warrants further work with improved measurement conditions. Nonetheless, the results strongly indicate that such n-type or p-type silicon wafers are candidate materials for use in self-cooling devices.     

Temperature renormalization of the conduction electron g factor in silicon

The temperature dependence of the conduction electron g factor in silicon is studied theoretically and experimentally. The theory of renormalization of the electron energy in an external magnetic field due to interaction with lattice vibrations is developed. According to the results of the calculation, in second-order perturbation theory, the temperature renormalization of the electron g factor is determined mainly by the difference between the processes of intervalley scattering that occur with spin preservation for spin-up and spin-down electrons. Experimental investigations of n-Si samples by the electron spin resonance method demonstrated the almost linear decrease in the conduction electron g factor with an increase in the temperature in a wide range from 80 to 300 K. The results of the calculation are in good agreement with the experimental data.     

A physical model for the dependence of carrier lifetime on doping density in nondegenerate silicon

A theoretical model that describes the dependence of carrier lifetime on doping density, which is based on the equilibrium solubility of a single defect in nondegenerately doped silicon, is developed. The model predictions are consistent with the longest measured hole and electron lifetimes reported for n-type and p-type silicon, and hence imply a possibly “fundamental” (unavoidable) defect in silicon. The defect is acceptor-type and is more soluble in n-type than in p-type silicon, which suggests a longer fundamental limit for electron lifetime than for hole lifetime at a given nondegenerate doping density. The prevalent, minimum density of the defect, which defines these limits, occurs at the processing temperature below which the defect is virtually immobile in the silicon lattice. The analysis reveals that this temperature is in the range 300–400°C, and thus emphasizes, when related also to common non-fundamental defects, the significance of low-temperature processing in the fabrication of silicon devices requiring long or well-controlled carrier lifetimes.