用于机器人轨迹定位纠正系统的回声状态神经网络研究

循环神经网络(RecurrentNeuralNetworks)是人工神经网络(ArtificialNeuralNetworks)中重要的分支,与前馈神经网络(ForwardNeuralNetworks)相比具有更好的时间序列学习能力。但长期以来其学习法一直不能脱离前馈神经网络而自成一体,回声状态神经网络(EchoStateNetworks(ESN))是打破这一局面的全新学习方法。其独特的结构,良好的短期记忆能力,方便的学习方法,不俗的非线性特性是以前循环神经网络所不可比的。本文在介绍了回声状态神经网络之后将其用于四轮机器人的位置测量系统中,有良好的表现。     

Preparation and properties of novel polyurethane insulating coatings based on glycerin‐terminated urethane prepolymers and blocked isocyanate

Novel polyurethane insulating coatings were prepared from the reaction of glycerin‐terminated polyurethane prepolymers (GPUPs) and a blocked isocyanate curing agent (BIC). The GPUPs were prepared from the reaction of one equivalent of polycaprolactone polyol (CAPA 210) with an excess amount of 4,4′‐methylene bis(phenyl isocyanate) (MDI) and subsequent reaction of the NCO‐terminated polyurethane with glycerin. The BIC was prepared from the reaction of trimethylol propane (TMP), toluene diisocyanate (TDI) and N‐methylaniline (NMA). The polyols and curing agent were characterized by conventional methods while the curing condition was optimized via gel content measurements. The curing kinetics of the polyurethane coating were investigated and the kinetic parameters derived. The crosslink densities of the samples were determined via the equilibrium swelling method, using the Flory–Rehner equation. The relationships between the crosslink density and the electrical, physical, mechanical and dynamic mechanical properties of the coatings were also studied. Copyright © 2005 Society of Chemical Industry     

Low density polyethylene and grafted lignin polyblends using epoxy‐functionalized compatibilizer: mechanical and thermal properties

Lignin was graft copolymerized with methyl methacrylate using manganic pyrophosphate as initiator. This modified lignin was then blended (up to 50 wt%) with low density polyethylene (LDPE) using a small quantity of poly[ethylene‐co‐(glycidyl methacrylate)] (PEGMA) compatibilizer. The mechanical properties of the blend were substantially improved by using modified lignin in contrast to untreated lignin. Differential scanning calorimetry studies showed loss of crystallinity of the LDPE phase owing to the interaction between the blend components. Thermogravimetric analysis showed higher thermal stability of modified lignin in the domain of blend processing. This suggested that there is scope for useful utilization of lignin, which could also lead to the development of eco‐friendly products. Copyright © 2005 Society of Chemical Industry     

Finite‐Element Implementation of a Self‐consistent Texture Model with a Hardening Law Based on Dislocation Densities

A self‐consistent texture model is implemented in the finite‐element code ABAQUS/Explicit®. Hardening of the slip systems is described by evolution laws for dislocation densities, and the Bauschinger effect is included. With these ingredients the cold rolling of an IF‐steel is simulated and the evolution of texture and anisotropy is examined. The use of the dislocation based hardening law is found to have a significant effect on the evolution of the anisotropic yield surface, but only very little influence on the texture. Furthermore, the hardening model is used to investigate the evolution of dislocation densities in differently oriented grains during plane strain compression. It is found that the dislocation density varies greatly depending on orientation, an observation which can help to explain the orientation dependence of recrystallization.     

三角函数型增长曲线预测油田开发指标

提出了用三角函数表示的一类用于油田开发指标预测的增长曲线共6条，从而进一步完善和丰富了现有的增长曲线法。提出和推导了它们的Np-t、Qt-t、Qt-Np关系式、（Np/NRmax)Qtmax准数计算式以及参数求解式。其特片研究表明，它们特征相近；（Np/NRamx)Qtmax变化范围为0-0.36785，属于峰值产量出现在油田开发初期的情况，其Qt-Np关系为一不对称的拱形曲线。提出了求取曲线参数的过原点重复线性回归法。双河油田的实例表明该方法有一定实用价值。     

A probabilistic model for cleavage fracture with a length scale--parameter estimation and predictions of stationary crack experiments

This study presents a large experimental investigation in the transition temperature region on a modified A508 steel. Tests were carried out on single-edge-notch-bend specimens with three different crack depth over specimen width ratios to capture the strong constraint effect on fracture toughness. Three test temperatures were considered, covering a range of 85 °C. All specimens failed by cleavage fracture prior to ductile tearing. A recently proposed probabilistic model for the cumulative failure by cleavage was applied to the comprehensive sets of experimental data. This modified weakest link model incorporates a length scale, which together with a threshold stress reduce the scatter in predicted toughness distributions as well as introduces a fracture toughness threshold value. Model parameters were estimated by a robust procedure, which is crucial in applications of probabilistic models to real structures. The conformity between predicted and experimental toughness distributions, respectively, were notable at all the test temperatures.     

Preparation of ultra‐low‐density microcellular materials

Microcellular polymeric materials can be obtained by the polymerization of a high‐internal‐phase emulsion. These materials are good candidates as targets toward inertial confinement fusion. This application requires severe specifications, including a very low density and a small cell size. In this study, we examined the influence of parameters such as emulsification conditions, surfactant nature, and the presence of a porogen on the obtainment of those requirements. It was possible to obtain microcellular polymeric foams with apparent densities as low as 0.0126 g/cm³. However, it was difficult to obtain very low material density and still maintain a small average pore size. © 2005 Wiley Periodicals, Inc. J Appl Polym Sci 96: 2053–2063, 2005     

从烟草悬浮细胞中分离液泡前体的方法研究

通过跟踪含有液泡前体特异标记物的组分, 利用细胞壁酶水解胞壁制造原生质体, 并用真针头挤压的方法裂解细胞, 用蔗糖密度梯度离心的方法分离液泡前体. 通过相差显微镜观察, 发现所分离的液泡前体保持完整的结构. 利用不同细胞器的标记物抗体进行免疫标记检测, 证明所纯化的液泡前体不含其他细胞器, 具有较高的纯度.     

市区微蜂窝传播路径损耗模式的比较检验

讨论了几种微蜂窝路径损耗模式,并用我国某城市测量的微蜂窝传播数据对COST231WI,修正COST231WI和Xia模式进行了检验,结果表明修正COST231模式的误差最小,Xia模式次之,COST231WI模式最差。从而得到修正COST231模式比较符合我国实际的微蜂窝传播预测模式。     

Robustness of classification rules that incorporate additional information

The discrimination problem for two normal populations with the same covariance matrix when additional information on the population is available is considered. A study of the robustness properties against training sample contamination of classification rules that incorporate this additional information is performed. These rules have received recently attention where their total misclassification probability (TMP) is proved to be lower than Fisher's linear discriminant rule. The results of a simulation study on the TMP which compares the behaviour of the new rules against Fisher's rule and some of its robustified versions under different types of contamination are presented. These results show that the rules that incorporate the additional information not only have lower TMP, but they also prevent against some types of contamination. In order to achieve prevention from all types of contamination a robustifed version of these rules is recommended.