Hybrid three-dimensional variation and particle filtering for nonlinear systems

This work addresses the problem of estimating the states of nonlinear dynamic systems with sparse observations.We present a hybrid three-dimensional variation(3DVar) and particle piltering(PF) method,which combines the advantages of 3DVar and particle-based filters.By minimizing the cost function,this approach will produce a better proposal distribution of the state.Afterwards the stochastic resampling step in standard PF can be avoided through a deterministic scheme.The simulation results show that the performance of the new method is superior to the traditional ensemble Kalman filtering(EnKF) and the standard PF,especially in highly nonlinear systems.     

Improved particle filters for multi-target tracking

We present a novel approach for improving particle filters for multi-target tracking. The suggested approach is based on drift homotopy for stochastic differential equations. Drift homotopy is used to design a Markov Chain Monte Carlo step which is appended to the particle filter and aims to bring the particle filter samples closer to the observations while at the same time respecting the target dynamics. We have used the proposed approach on the problem of multi-target tracking with a nonlinear observation model. The numerical results show that the suggested approach can improve significantly the performance of a particle filter.     

基于稀疏贝叶斯学习的复杂网络拓扑估计

提出了一种噪声环境下复杂网络拓扑估计方法, 仅利用含噪时间序列估计未知结构混沌系统的动力学方程和参数, 以及由混沌系统组成的复杂网络的拓扑结构、节点动力学方程、所有参数、 节点间耦合方向和耦合强度.通过采用动力学方程的统一形式, 将动力系统方程结构和参数估计看成线性回归问题的系数估计, 该估计问题利用贝叶斯压缩传感的信号重建算法求解, 含噪信号的模型重建使用相关向量机方法,即通过稀疏贝叶斯学习求解稀疏欠定线性方程得到上面提到的可估计对象.以单个Lorenz系统及由200个 Lorenz系统组成的无标度网络为例说明方法的有效性. 仿真结果表明,提出的方法对噪声有很强的鲁棒性,收敛速度快,稳态误差极小, 克服了最小二乘估计方法收敛速度慢、 稳态误差大以及压缩传感估计方法对噪声鲁棒性不强的缺点.     

基于粒子群算法的稀疏阵列超声相控阵全聚焦成像

为降低相控阵超声检测全聚焦算法的成像数据量及阵列稀疏优化的计算时间,研究了一种用于稀疏阵列全聚焦成像的阵列优化算法,并通过实验对其成像效果进行了验证。针对目前超声相控阵检测的全矩阵采集数据量大、全聚焦算法成像时间长的难点,该文通过构建稀疏阵列,在保证成像质量的同时显著降低成像数据量,提高了全聚焦算法的成像效率。通过以主瓣宽度、旁瓣峰值以及主瓣峰值作为约束条件构建适应度函数,采用粒子群算法得到稀疏阵元位置分布并进行阵元权重修正,并将其用于稀疏全聚焦成像。相比全阵元成像,使用粒子群算法所得的稀疏阵列的阵元个数降低了56.25%、65.62%,数据使用量降低了80.86%、88.18%。在阵列优化方面,相比遗传算法减少了84.86%的计算时间。     

Kinetic Description of a Rayleigh Gas with Annihilation

In this paper, we consider the dynamics of a tagged point particle in a gas of moving hard-spheres that are non-interacting among each other. This model is known as the ideal Rayleigh gas. We add to this model the possibility of annihilation (ideal Rayleigh gas with annihilation), requiring that each obstacle is either annihilating or elastic, which determines whether the tagged particle is elastically reflected or removed from the system. We provide a rigorous derivation of a linear Boltzmann equation with annihilation from this particle model in the Boltzmann–Grad limit. Moreover, we give explicit estimates for the error in the kinetic limit by estimating the contributions of the configurations which prevent the Markovianity. The estimates show that the system can be approximated by the Boltzmann equation on an algebraically long time scale in the scaling parameter.

     

On the Blow Up and Condensation of Supercritical Solutions of the Nordheim Equation for Bosons

In this paper we prove that the solutions of the isotropic, spatially homogeneous Nordheim equation for bosons with bounded initial data blow up in finite time in the L ^∞ norm if the values of the energy and particle density are in the range of values where the corresponding equilibria contain a Dirac mass. We also prove that, in the weak solutions, whose initial data are measures with values of particle and energy densities satisfying the previous condition, a Dirac measure at the origin forms in finite time.     

Approximate importance sampling Monte Carlo for data assimilation

Importance sampling Monte Carlo offers powerful approaches to approximating Bayesian updating in sequential problems. Specific classes of such approaches are known as particle filters. These procedures rely on the simulation of samples or ensembles of the unknown quantities and the calculation of associated weights for the ensemble members. As time evolves and/or when applied in high-dimensional settings, such as those of interest in many data assimilation problems, these weights typically display undesirable features. The key difficulty involves a collapse toward approximate distributions concentrating virtually all of their probability on an implausibly few ensemble members.

After reviewing ensembling, Monte Carlo, importance sampling and particle filters, we present some approximations intended to moderate the problem of collapsing weights. The motivations for these suggestions are combinations of (i) the idea that key dynamical behavior in many systems actually takes place on a low dimensional manifold, and (ii) notions of statistical dimension reduction. We illustrate our suggestions in a problem of inference for ocean surface winds and atmospheric pressure. Real observational data are used.     

Accumulation of particles in time-dependent thermocapillary flow in a liquid bridge. Modeling of experiments

The study addresses the phenomenon of accumulation of rigid tracer particles suspended in a time-dependent thermocapillary flow in a liquid bridge. We report the results of the three-dimensional numerical modeling of recent experiments [1,2] in a non-isothermal liquid column. Exact physical properties of both liquids and particles are used for the modeling. Two liquids are investigated: sodium nitrate (NaNO₃) and n-decane (C₁₀H₂₂). The particles are modeled as perfect spheres suspended in already well developed time-dependent thermocapillary flow. The particle dynamics is described by the Maxey-Riley equation. The results of our simulations are in excellent agreement with the experimental observations. For the first time we reproduced numerically formation of the particle accumulation structure (PAS) both under gravity and under weightlessness conditions. Our analysis confirms the experimental observations that the existence of PAS depends on the strength of the flow field, on the ratio between liquid and particle density, and on the particle size.     

Stability of the square-wave detonation in a model system

Using a set of model equations for reactive flow, we study the stability of a “square-wave” detonation, in which each particle of the fluid reacts instantaneously after an induction time which depends on how hard it was shocked. We obtain a differential-difference equation for the shock velocity, valid for small perturbations about the steady solution. This equation is of so-called “advanced” type, in which the velocity at a given time depends on both velocity and acceleration at an earlier time.     

Elliptic random-walk equation for suspension and tracer transport in porous media

We propose a new approach to transport of the suspensions and tracers in porous media. The approach is based on a modified version of the continuous time random walk (CTRW) theory. In the framework of this theory we derive an elliptic transport equation. The new equation contains the time and the mixed dispersion terms expressing the dispersion of the particle time steps. The properties of the new equation are studied and the fundamental analytical solutions are obtained. The solution of the pulse injection problem describing a common tracer injection experiment is studied in greater detail. The new theory predicts delay of the maximum of the tracer, compared to the velocity of the flow, while its forward “tail” contains much more particles than in the solution of the classical parabolic (advection-dispersion) equation. This is in agreement with the experimental observations and predictions of the CTRW theory.     

High energy description of dark energy in an approximate 3-brane Brans–Dicke cosmology

We consider a Brans–Dicke cosmology in five-dimensional space–time. Neglecting the quadratic and the mixed Brans–Dicke terms in the Einstein equation, we derive a modified wave equation of the Brans–Dicke field. We show that, at high energy limit, the 3-brane Brans–Dicke cosmology could be described as the standard one by changing the equation of state. Finally as an illustration of the purpose, we show that the dark energy component of the universe agrees with the observations data.     

Ray tracing/correlation approach to estimation of surface-based duct parameters from radar clutter

This paper describes a technique to estimate surface-based duct parameters by using a simple ray tracing/correlation method. The approach is novel in that it incorporates the Spearman rank-order correlation scheme between the observed surface clutter and the surface ray density for a given propagation path. The simulation results and the real data results both demonstrate the ability of this method to estimate surface-based duct parameters. Compared with the results obtained by a modified genetic algorithm combined with the parabolic wave equation, the results retrieved from the ray tracing/correlation scheme show a minor reduction in accuracy but a great improvement on computation time. Therefore the ray tracing/correlation method might be used as a precursor to more sophisticated and slower techniques, such as genetic algorithm and particle filters, by narrowing the parameter search space and providing a comprehensive and more efficient estimation algorithm.     

The oscillation of stock price by majority orienting traders with investment position

We consider an interacting particle system for the stock price fluctuation. The change of the stock price with a feedback by the price considering the herding behavior (majority orienting behavior) of traders, gives the van der Pol equation as a deterministic approximation. Considering the investment position of each trader, we introduce the delayed van der Pol equation. The history of investment positions, for example sell or buy, of each trader for a stock makes a memory effect, which is modeled by using the time retardation. The delayed van der Pol equation model seems to be natural and explains typical phenomena, for example triangle pattern, volatility jumps, price jumps and price trends, known for the time series of a stock price.     

利用杜哈梅尔方法近似求解EDF的自发辐射特性

掺铒光纤(EDF)中的自发辐射噪声是影响掺铒光纤放大器(EDFA)工作性能和掺铒光纤激光器(EDFL)的起振特性的重要因素。自发辐射与泵浦方式紧密相关,研究脉冲泵浦下EDF的自发辐射具有重要的学术意义。从速率方程出发,建立了任意波形脉冲泵浦下EDF自发辐射的能级粒子数分布所满足的一元二阶变系数微分方程。由于没有封闭形式的解析解,采用杜哈梅尔方法,将泵浦脉冲波形进行分时段描述,每个小时段都有解析解,从而得到了自发辐射的平均功率表达式。分析结果表明,随着泵浦功率的增大,ASE的输出波形更接近泵浦光的波形,且泵浦光的毛刺对于ASE噪声的影响较小。     

时变偏微分方程的贝叶斯稀疏识别方法

在数据驱动的建模中,通过测量或模拟得到时空数据,我们发现基于拉普拉斯先验的贝叶斯稀疏识别方法能有效地恢复时变偏微分方程的稀疏系数.本文将贝叶斯稀疏识别方法运用于各种时变偏微分方程模型(KdV方程、Burgers方程、Kuramoto-Sivashinsky方程、反应-扩散方程、非线性薛定谔方程和纳维-斯托克斯方程)的方...     

On the variable order dynamics of the nonlinear wake caused by a sedimenting particle

In this work we develop a variable order (VO) differential equation of motion for a spherical particle sedimenting in a quiescent viscous liquid. In particular, we examine the various force terms in the equation of motion and propose a new form for the history drag acting on the particle. We show that the variable order formulation allows for an effective way to express the dynamic transition of the dominant forces over the entire time of the motion of the particle from rest to terminal velocity. The use of VO operators also allows us to examine the evolving dynamics of the wake during sedimentation. Using numerical data from a finite element simulation of a sedimenting particle, we first solve for the order of the derivative that returns the correct decay of the history force. We then propose a relatively simple expression for the history force that is a function of the Reynolds number and particle-to-fluid density ratio. The new history drag expression correlates very well (R²>0.99) with the numerical data for terminal Reynolds numbers ranging from 2.5 to 20, and for particle-to-fluid density ratios of interest in practice (1<β<10).     

Equations for the distributions of functionals of a random-walk trajectory in an inhomogeneous medium

Based on the random-trap model and using the mean-field approximation, we derive an equation that allows the distribution of a functional of the trajectory of a particle making random walks over inhomogeneous-lattice site to be calculated. The derived equation is a generalization of the Feynman-Kac equation to an inhomogeneous medium. We also derive a backward equation in which not the final position of the particle but its position at the initial time is used as an independent variable. As an example of applying the derived equations, we consider the one-dimensional problem of calculating the first-passage time distribution. We show that the average first-passage times for homogeneous and inhomogeneous media with identical diffusion coefficients coincide, but the variance of the distribution for an inhomogeneous medium can be many times larger than that for a homogeneous one.     

One-dimensional fluid and circuit simulation of an AC plasmadisplay cell

A one-dimensional multispecies model of a helium filled AC plasma display cell is described. The model includes a continuity equation for each species. Poisson's equation and a circuit equation. The fill gas is helium for which the cross sections for ionization and excitation for the various atomic states are well known. The reaction rates for the different species were parametrized as functions of E/N using a zero-dimensional spatial and two velocity multispecies Boltzmann kinetic code. Calculations are performed using address and sustain voltages operating at a frequency of 50 kHz to model the pulse behavior of a dielectric barrier discharge which occurs over a time scale on the order of 200 ns or less. The code simulates multipulse behavior and shows reasonably good agreement with experimental data. The simulations show that the circuit filters out the discharge dynamics and that the pulse width of the discharge current depends on the circuit RC time constant. Using a 123 V sustain voltage, the calculated discharge current is 0.4 A with a full width at half maximum (FWHM) of 28 ns. Experimental values using a 120 V sustain voltage are 0.32 A discharge current and 42 ns FWHM     

Scattering of electromagnetic waves by ensembles of particles and discrete random media

Current problems of the theory of multiple scattering of electromagnetic waves by discrete random media are reviewed, with an emphasis on densely packed media. All equations presented are based on the rigorous theory of electromagnetic scattering by an arbitrary system of non-spherical particles. The main relations are derived in the circular-polarization basis. By applying methods of statistical electromagnetics to a discrete random medium in the form of a plane-parallel layer, we transform these relations into equations describing the average (coherent) field and equations for the sums of ladder and cyclical diagrams in the framework of the quasi-crystalline approximation. The equation for the average field yields analytical expressions for the generalized Lorentz-Lorenz law and the generalized Ewald-Oseen extinction theorem, which are traditionally used for the calculation of the effective refractive index. By assuming that the particles are in the far-field zones of each other, we transform all equations asymptotically into the well-known equations for sparse media. Specifically, the equation for the sum of the ladder diagrams is reduced to the classical vector radiative transfer equation. We present a simple approximate solution of the equation describing the weak localization (WL) effect (i.e., the sum of cyclical diagrams) and validate it by using experimental and numerically exact theoretical data. Examples of the characteristics of WL as functions of the physical properties of a particulate medium are given. The applicability of the interference concept of WL to densely packed media is discussed using results of numerically exact computer solutions of the macroscopic Maxwell equations for large ensembles of spherical particles. These results show that theoretical predictions for spare media composed of non-absorbing or weakly absorbing particles are reasonably accurate if the particle packing density is less than ∼30%. However, a further increase of the packing density and/or absorption may cause optical effects not predicted by the low-density theory and caused by near-field effects. The origin of the near-filed effects is discussed in detail.     

激光点火煤粒着火的模型分析

本文对颗粒煤在激光加热条件下的着火和燃烧进行了数值模拟。采用的是一个简单的煤粒着火与燃烧的一维模型。该模型采用了热解和双平行反应模型,考虑了煤粒表面的多相反应和气相的基元反应以及气相中的传热与传质。从获得的煤粒表面和气相空间的温度随时间的变化规律,可以判断不同煤种的着火方式。