HiSIM2 Circuit simulation - Solving the speed versus accuracy crisis

The development trend in compact modeling goes toward surface-potential-based approaches and leads to models like HiSIM2, with higher accuracy, fewer model parameters, and shorter computer runtime than achievable with the conventional threshold-voltage-based approaches. The main motivation for continuing this development effort is to realize a sufficient design capability of RF circuits with advanced MOSFETs, where many higher-order phenomena affect the circuit performance, as well as of large mixed-signal circuits, where both accuracy and short simulation time are a must. The trend toward the surface potential brings compact modeling for circuit simulation also much closer to 2D and three-dimensional numerical device simulation. Therefore, both approaches can now come together and work united to achieve the common goal of realizing rapid technology progress for the benefit of the society     

一种蒸发波导预测的改进模型

利用大概率的近海面蒸发波导效应实现雷达超视距探测已成为目前岸基、舰船无线电测量系统的关键技术之一。在实际应用中,需要由蒸发波导参数建立大气修正折射率剖面的预测模型。针对目前常用的蒸发波导预测只含波导高度和海面修正折射率两个参数,没有考虑波导强度的影响,使得该模型存在一定的误差的现状,通过在原有模型的基础上增加波导强度的影响,建立了高精度的改进模型,使得蒸发波导预测模型的精度得到了有效提高,进而可进一步提高雷达超视距探测精度。实验验证表明,蒸发波导预测改进模型的精度远远高于常用的模型,且波导强度越强,则改正模型的精度越高。     

On the accuracy and efficiency of substrate current calculationsfor sub-μm n-MOSFET's

The accuracy and efficiency of the self-consistent (regarding the electric field) Monte Carlo model, nonself-consistent Monte Carlo model, and the soft-threshold lucky electron model (LEM) for the calculation of substrate currents in deep sub-μm n-MOSFET's are investigated. While the two Monte Carlo models are in good agreement with the experiment, the simpler LEM model still gives reasonable results even for a 0.16 μm n-MOSFET. On the other hand, huge differences in the CPU time consumption are found and the LEM is about four orders of magnitude faster than the self-consistent Monte Carlo simulations. The nonself-consistent calculations are only one order of magnitude slower than the LEM. The good agreement with the experiment is obtained without considering the so-called surface impact ionization or any fitting of parameters on the device level     

非掺杂对称双栅的基于完整表面电势的核心模型

通过求解Poisson方程自洽地得到了表面电势随沟道电压的变化关系,从而推出了非掺杂对称双栅MOSFET的一个基于表面势的模型.通过Pao-Sah积分得到了漏电流的表达式.该模型由一组表面势方程组成,解析形式的漏电流可以通过源端和漏端的电势得到.结果标明该模型在双栅MOSFET的所有工作区域都成立,而且不需要任何简化(如应用薄层电荷近似)和辅助拟合函数.对不同工作条件和不同尺寸器件的二维数值模拟与模型的比较进一步验证了提出模型的精度.     

A simple two-dimensional model for IGFET operation in the saturation region

A model is developed for an IGFET operating in saturation and accounting for the two-dimensional potential distribution in the section of the surface space-charge region adjacent to the drain. This section is treated as a volume obeying Gauss's law, thereby enabling the charge contained in it to be related to the integral of the electric displacement density normal to its surface without the need to consider the detailed distribution of charge inside. The resulting model shows, explicitly, the dependence of device output characteristics on applied potentials, geometric and processing parameters. It is also shown that by making adequate approximations, simple yet accurate forms of the model are obtained. The accuracy of the model is demonstrated by comparisons between calculated and measured device output characteristics.     

A Physics-Based Analytic Solution to the MOSFET Surface Potential From Accumulation to Strong-Inversion Region

A physics-based analytic solution to the surface potential from the accumulation to the strong-inversion region has been derived from the complete MOSFET surface potential equation in this paper without any need for smooth functions or simplification by dropping some second-order related terms. Its high accuracy in predicting the surface potential and the transcapacitance under various bias conditions has also been verified by a comparison with the numerical results. The explicit surface-potential solution not only leads to a more clear understanding of MOSFET device physics but also provides a better platform to develop the advanced surface potential-based model for the circuit simulation.     

基于改进CenterNet的AFP铺层表面缺陷检测

针对利用可见光图像检测AFP铺层表面缺陷受光源条件差、预浸纱纹理对比度低等因素影响,检测结果不理想,提出一种基于改进CenterNet的AFP铺层表面红外图像缺陷检测方法,提高AFP铺层表面缺陷检测性能。首先,针对CenterNet模型参数数量过多而工控机硬件配置有限的问题,提出利用基于ASFF的轻量级MobileNetV3作为骨干网络,构建轻量级anchor-free检测模型AFP-CenterNet,减少网络参数数量的同时降低计算机存储资源占用率。然后,针对高斯核函数带宽参数的求解,提出一种根据ground-truth bounding box长宽比自适应调整带宽参数的方法,减小负样本数量,降低网络模型的损失误差。实验结果表明,改进后的AFP-CenterNet在AFP红外数据集上的AP为90.2%,模型内存容量为12.9 MB,使用GPU加速时单张检测时间为52 ms。和原有的CenterNet骨干网络相比,AFP-CenterNet检测精度略差于DLA-34,和ResNet-101相当,比ResNet-18高7.7%,内存占用率和DLA-34、ResNet-101、ResNet-18相比分别降低83.2%、93.6%和78.6%。和SSD、YOLOv3相比,AFP-CenterNet模型的AP分别提升9.6%和8.3%,内存占用量降低85.1%和94.5%。在不使用GPU加速的条件下,改进后的AFP-CenterNet的检测速度和CenterNet、SSD、YOLOv3相比提高近一倍,具有明显的检测优势。     

一种新的星载SAR图像定位方法的研究

该文应用摄影测量理论,将星载SAR图像近似看作一个画幅式投影传感器获取的等效影像,从而对星载 SAR共线方程进行改进,提出了一种新的定位模型。文中选用一景北京地区RadarSat影像进行实验,结果表明,该模型表达简单,解算稳定,不需要已知轨道信息和成像参数且定位精度保持在像素级范围之内。     

A rectilinear Gaussian model for estimating straight-line parameters

For characterizing straight lines in defocused images, a rectilinear Gaussian model (RGM) is proposed. Based on this model, a novel method for estimating the parameters of straight lines is presented. This method, called gray-scale least square (GLS) method, directly deals with gray-scale image data without requiring any preprocessing and hence no additional noise is introduced. Furthermore, the method is able to simultaneously estimate four parameters of straight lines by performing the algorithm only once, while two parameters can be typically estimated by traditional method. Besides this, all parameters are given in closed-form solution. In order to illustrate the effectiveness of RGM and the GLS method, the experiments are performed on a set of artificial images and natural images. The experimental results show that the GLS method outperforms the traditional method from the point of view of sensitivity to noise and accuracy of parameter estimation.     

Simple algorithm for soil moisture retrieval with co-polarized SAR data

In this paper, an empirical methodology to retrieve bare soil moisture by Synthetic Aperture Radar (SAR) is developed. The model is based on Advanced Integral Equation Model (AIEM). Since AIEM cannot express cross-polarized backscattering coefficients accurately, we propose an empirical model to retrieve soil moisture for bare farmland only with co-polarized SAR data. The soil moisture can be obtained by solving an equation of HH and VV polarized data without any field measurements. Both simulated and real SAR data are used to validate the accuracy of the model. This method is especially effective in a large area where the surface roughness is difficult to be completely measured.     

SiGe HBT小信号等效电路的参数直接提取

提出了一种求解硅锗异质结双极型晶体管(SiGe HBT)小信号等效电路模型的参数直接提取方法.整个提取过程使用由小信号等效电路推导出的一系列解析表达式,不使用任何数值优化方法.参数提取结果使用ADS软件仿真验证.结果表明,该方法简单易行,较为精确.     

基于GoogLeNet的手写体汉字识别

近年来基于深度学习的方法识别手写体汉字取得了很多突破,但现有的一些方法存在计算参数多、模型收敛慢、训练时间长的缺点。针对以上问题,提出了基于GoogLeNet的脱机手写体汉字识别模型HCCR-IncBN,模型使用了5个Inception-v2模块,训练参数较少,模型收敛更快,存储整个模型只需要26MB的存储空间。实验利用HCCR-IncBN模型在ICDAR2013数据集获得了95.94%的识别准确率,表明模型在没有使用任何手写体汉字的特定领域知识和无需人工提取其他特征的前提下能够获得较高的识别效果。     

Reflow of supported sub-100 nm polymer films as a characterization process for NanoImprint lithography

In this paper, we address the reflow behavior of polymer thin films, focusing our effort on the accuracy of surface shape recognition. Although much work was already performed to control resist reflow during lens manufacturing for instance, our approach is significantly different since no contact line (substrate/polymer/atmosphere) needs to be considered. Here, a linear stability approach is successfully developed to describe the thin film evolution which is also compared with experiments. Polystyrene films, with thickness ranging from few tens of nanometers up to several hundred of nanometers were patterned with NanoImprint lithography technique. Atomic force microscopy measurements were used to characterize smooth or steep shapes, respectively. Mechanical measurements of earlier stages of pattern reflow were directly accessible without any assumption, contrary to the diffraction method usually employed. We show that by controlling the reflow process of any complex surface shape during the course of time, measurements of material parameters such as thin film viscosity, surface tension, or even Hamaker constant can be made possible.     

基于椭圆轨道的Geo-SAR精确多普勒参数解析计算方法

该文综合考虑了轨道偏心率、地球扁率和地球自转三方面的因素,推导得出高精度星载SAR多普勒中心频率和多普勒调频率解析式。鉴于2维姿态导引对于Geo-SAR的必要性,进一步分析了姿态导引后和零姿态导引两种情况下的多普勒参数,为Geo-SAR的多普勒参数计算提供了实用且有效的方法。该文得到的多普勒参数解析计算式适用于任意轨道高度和任意姿态的星载SAR,且具备很高的精度。基于该结论,研究了运行于椭圆轨道的Geo-SAR多普勒中心和多普勒调频率特性。     

Automatic variogram parameter extraction for textural classification of the panchromatic IKONOS imagery

Range and sill are two important parameters of a variogram. Their extraction usually involves experimental fitting of variograms using models specified by the analyst and requires much use of trial and error. The objective of this paper is to design an algorithm for extracting the range and sill of a variogram automatically without fitting a model. Combined with the semivariance at the lag of one pixel (/spl gamma//sub 1/), the extracted range and sill were applied to the textural classification of a panchromatic IKONOS image over Xichang, Sichuan Province, China. Results show that any of these three parameters can lead to the increase of the classification accuracy. When all three parameters were used with the raw image data, the average kappa statistic for five window sizes increased from 0.24 to 0.76, indicating promise of the range and sill in texture classification.     

Parameter sensitivity of soil moisture retrievals from airborne C- and X-band radiometer measurements in SMEX02

Among passive microwave frequencies, sensors operating at C- and X-band frequencies have been used with some success to estimate near-surface soil moisture from aircraft and satellite platforms. The objective of this paper is to quantify the sensitivities of soil moisture retrieved via a single-channel single-polarization algorithm to the observed brightness temperature and to retrieval algorithm parameters of surface roughness, vegetation B parameter, and single-scattering albedo. Examination of the regions within the parameter space that produce accurate soil moisture retrievals reveals that reasonably accurate retrievals can be made over a range of conditions using a fixed set of input parameters. Retrievals with horizontally polarized brightness temperature observations are more consistent than with vertically polarized observations. At horizontal polarization, sensitivity to the input parameters is much greater for wet soils than for dry soils, whereas for vertical polarization the moisture dependence is much weaker. At vertical polarization, sensitivities to variations in all parameters are much lower. To ensure that retrieval accuracy specifications are consistently met, high soil moisture conditions should be used in defining parameter accuracy requirements. Given the spatial and temporal variability of vegetation and soil conditions, it seems unlikely that, for regions with substantial rapidly growing vegetation, the accuracy requirements for model parameters in a single-frequency, single-polarization retrieval algorithm can be met with current satellite products. For such conditions, any soil moisture retrieval algorithm using parameterizations similar to those of this study may require multiple frequencies, polarizations, or look angles to produce stable, reliable soil moisture estimates.     

Parameter sensitivity of soil moisture retrievals from airborne L-band radiometer measurements in SMEX02

Over the past two decades, successful estimation of soil moisture has been accomplished using L-band microwave radiometer data. However, remaining uncertainties related to surface roughness and the absorption, scattering, and emission by vegetation must be resolved before soil moisture retrieval algorithms can be applied with known and acceptable accuracy using satellite observations. Surface characteristics are highly variable in space and time, and there has been little effort made to determine the parameter estimation accuracies required to meet a given soil moisture retrieval accuracy specification. This study quantifies the sensitivities of soil moisture retrieved using an L-band single-polarization algorithm to three land surface parameters for corn and soybean sites in Iowa, United States. Model sensitivity to the input parameters was found to be much greater when soil moisture is high. For even moderately wet soils, extremely high sensitivity of retrieved soil moisture to some model parameters for corn and soybeans caused the retrievals to be unstable. Parameter accuracies required for consistent estimation of soil moisture in mixed agricultural areas within retrieval algorithm specifications are estimated. Given the spatial and temporal variability of vegetation and soil conditions for agricultural regions it seems unlikely that, for the single-frequency, single-polarization retrieval algorithm used in this analysis, the parameter accuracy requirements can be met with current satellite-based land surface products. We conclude that for regions with substantial vegetation, particularly where the vegetation is changing rapidly, any soil moisture retrieval algorithm that is based on the physics and parameterizations used in this study will require multiple frequencies, polarizations, or look angles to produce stable, reliable soil moisture estimates.     

Model-based estimation of wind fields over the ocean from windscatterometer measurements. I. Development of the wind field model

For pt.II see ibid., vol.28, no.3, p.361-373 (1990). A parametric model for near-surface mesoscale wind fields suitable for use in a model-based approach to estimate the wind field over a large region of the scatterometer measurement swath from measurements of the radar backscatter of the ocean's surface is discussed. The assumptions made in developing the model are described, and the accuracy of the resulting model is evaluated. The model is based only on scatterometer data and is computationally tractable, consisting of a linear equation relating the near-surface wind field to a vector of model parameters which is estimated from the scatterometer measurements. Tradeoffs in the accuracy of the model and the number of unknown parameters are considered     

DC thermal modelling of GaAs MESFETs based on a semiempirical approach

In this paper a new semiempirical large signal thermal model of GaAs MESFETs is proposed for DC characterization. The model includes a third order dependence of fitting parameters on bias conditions and three thermal fitting parameters. A number of GaAs MESFETs, each very different from a geometrical and technological point of view, have been characterized as a function of temperature and modelled with high accuracy. The CPU extraction time results are moderate in any example. Results have been compared with the Rodriguez–Tellez model, showing improvements in accuracy of better than 30%. The model can be successfully used in MMIC CAD applications.