聚焦形貌恢复方法模型设计

采用基于拉普拉斯算符聚焦形貌恢复方法,提出了模拟目标深度测量的数值模型。数值模拟的核心是基于通过几何光学预测的理想图像的卷积与透镜广义孔径函数的多色点扩散函数,即用聚焦误差替代抛物线圆柱形貌或高斯函数。该模型可以使用基于聚焦形貌恢复方法的传感器真实组件参数、光源光谱、光学系统离差、相机的光谱灵敏度。提出了光学系统离差（消球差、消色差、色差）对确定目标表面形貌的精确度和可靠性的影响。结果表明,该模型可以有效提高实验效率,缩短时滞,降低成本。     

Zero Modes’ Fusion Ring and Braid Group Representations for the Extended Chiral <Emphasis Type="Italic">su</Emphasis>(2) WZNW Model

A zero modes’ Fock space is constructed for the extended chiral WZNW model. It gives room to a realization of the fusion ring of representations of the restricted quantum universal enveloping algebra at an even root of unity, and of its infinite dimensional extension by the Lusztig operators We provide a streamlined derivation of the characteristic equation for the Casimir invariant from the defining relations of A central result is the characterization of the Grothendieck ring of both and in Theorem 3.1. The properties of the fusion ring in are related to the braiding properties of correlation functions of primary fields of the conformal current algebra model.      

同型二聚体肽模型的二维红外光谱

利用一个同型二聚体模型,甲酰胺二聚体,模拟一对肽基团.用量子化学从头计算考察了二聚体中酰胺-I带振动模式之间的振动耦合及其在空间的行为. 研究发现C=O伸缩振动耦合以静电作用为主,耦合有效距离能超过10 oA. 一维和二维红外光谱的激子模拟计算表明,耦合常数的空间依赖性能够清楚地表现在光谱特征中. 这些结果意味着多肽中C=O伸缩振动模式能够在很远的距离相互耦合并产生振动态的离域化.     

Electrophoresis and electric conduction in a salt-free suspension of charged particles

The electrophoresis and electric conduction of a suspension of charged spherical particles in a salt-free solution are analyzed by using a unit cell model. The linearized Poisson-Boltzmann equation (valid for the cases of relatively low surface charge density or high volume fraction of the particles) and Laplace equation are solved for the equilibrium electric potential profile and its perturbation caused by the imposed electric field, respectively, in the fluid containing the counterions only around the particle, and the ionic continuity equation and modified Stokes equations are solved for the electrochemical potential energy and fluid flow fields, respectively. Explicit analytical formulas for the electrophoretic mobility of the particles and effective electric conductivity of the suspension are obtained, and the particle interaction effects on these transport properties are significant and interesting. The scaled zeta potential, electrophoretic mobility, and effective electric conductivity increase monotonically with an increase in the scaled surface charge density of the particles and in general decrease with an increase in the particle volume fraction, keeping each other parameter unchanged. Under the Debye-Hückel approximation, the dependence of the electrophoretic mobility normalized with the surface charge density on the ratio of the particle radius to the Debye screening length and particle volume fraction in a salt-free suspension is same as that in a salt-containing suspension, but the variation of the effective electric conductivity with the particle volume fraction in a salt-free suspension is found to be quite different from that in a suspension containing added electrolyte.     

质子水合结构的振动态密度分析

使用基于多态经验价键模型的分子动力学模拟, 对水溶液中质子的水合结构及其在质子传递过程中的动力学过程进行了研究. 在价键模型的方法下, 质子的水合结构主要以H₉O₄⁺(Eigen)以及过渡态的H₅O₂⁺(Zundel)结构形态存在, 且在这两种结构中以Eigen的形态表现明显. 通过对质子传递过程中不同水合结构的态密度频谱分析, 发现一个在2000~3000 cm^-1范围内的明显连续的宽吸收谱带, 主要归因于Eigen结构的贡献, 这些特征峰的出现与水合氢离子第一溶剂化层内的强氢键作用密切相关. 对于Zundel的结构, 在1760 cm^-1处出现一个较为明显的肩峰, 归属为质子传递模式的特征振动. 通过对质子水合结构态密度频谱的分析, 可望增强对于稀酸溶液红外光谱中的连续宽吸收带以及质子传递的微观动力学过程的理解.     

A domain decomposition method for three species modeling of multi-electrode negative corona discharge – With applications to electrostatic precipitators

The negative corona discharge problem for multi-electrode geometries is modeled by a three species model. The equations are solved using domain decomposition, by recognizing that multiple species are only present in a small part of the domain, and hence only need to be incorporated locally in the model.The method is applied to an industrially relevant three wire electrostatic precipitator geometry. The calculated current density is in good agreement with experimental data. To further illustrate applications of the three species solution, it is used for coupled particle, fluid, and electrostatic simulations to analyze particle collection properties.     

基于语义的软件服务故障自动诊断模型

模仿学习是机器人仿生机制研究的主要内容之一,即通过观察、理解、学习、模仿示教行为实现机器人的仿生特性。基于高斯过程分别表达采集离散示教信号所构成的示教轨迹和含有未知参数策略的模仿轨迹,构建模仿学习方法框架,将概率模型匹配引入到模仿学习中,以KL散度为代价函数比较两种轨迹的概率分布,运用梯度下降法寻求使KL散度最小的最优模仿控制策略,将策略应用于模仿机器人以完成与示教相同的模仿任务。以关节型机器人的机械臂摆动行为模仿为学习任务进行仿真,结果表明基于概率轨迹匹配的模仿学习方法能够实现机械臂摆动行为模仿,学习过程较传统方法简易且学习效果较好。     

基于Demons算法改进的图像去噪模型研究

在Demons算法的基础上, 将扩散过程看作图像配准, 建立一种新的基于图像配准的Demons 去噪模型. 实验表明, 该模型去噪效果优于经典的Perona-Malik模型, 排除了模型的病态性. 考虑到新模型在图像去噪过程中仅靠梯度信息表示图像的局部特征还不完善, 故将水平集曲率作为控制图像结构的驱动力因素引入到此模型中, 提出了一种新的梯度和曲率双重驱动力的图像去噪模型. 分析和仿真结果表明, 两种新模型都可有效抑制噪声, 清晰度也有明显的提高, 其中双重驱动力的图像去噪模型去噪效果更具优越性.