基于星凸形随机超曲面模型多扩展目标多伯努利滤波器

针对复杂不确定性环境下具有不规则形状的多扩展目标跟踪问题, 提出了一种基于星凸形随机超曲面模型(Star-convex RHM)的多扩展目标多伯努利滤波算法.首先, 在有限集统计(Finite set statistics, FISST)理论框架下, 采用多伯努利随机有限集(MBer-RFS)和泊松RFS (Possion-RFS)分别描述多扩展目标的状态和观测, 并给出扩展目标势均衡多目标多伯努利(ET-CBMeMBer)滤波器.其次, 利用RHM去描述任意星凸形扩展目标的量测源分布, 提出了容积卡尔曼高斯混合星凸形多扩展目标多伯努利滤波器.此外, 本文给出了一种多扩展目标不规则形状估计性能的评价指标.最后, 通过多扩展目标和具有形状突变的多群目标的跟踪仿真实验验证了本文方法的有效性.     

星凸形多扩展目标跟踪中的传感器控制方法

针对多扩展目标跟踪中的传感器控制问题, 本文基于有限集统计(FISST)理论与随机超曲面模型(RHM), 利 用多伯努利(MBer)滤波器提出有效的传感器控制策略. 首先, 文中给出多扩展目标跟踪中基于信息论联合目标形状 估计优化和目标运动状态估计优化的传感器控制方法的求解思路. 其次, 给出RHM容积卡尔曼高斯混合(GM)势均 衡多扩展目标多伯努利滤波算法的具体实现过程. 然后, 结合GM密度间的柯西施瓦兹(Cauchy-Schwarz)散度提出 相应的传感器控制决策方法. 此外, 详细推导了扩展目标势的后验期望(PENET)的GM实现, 并提出以GM–PENET 为评价函数的传感器控制方法. 最后, 通过构造随机星凸形多扩展目标的跟踪优化仿真实验验证了本文所提传感 器控制方法的有效性.     

基于GLMB滤波和Gibbs采样的多扩展目标有限混合建模与跟踪算法

本文针对杂波条件下多扩展目标的状态估计, 目标个数估计, 扩展目标形状估计问题, 提出了一种基于标签随机有限集(Labelled random finite sets, L-RFS)框架下多扩展目标跟踪学习算法, 该学习算法主要包括两方面:多扩展目标动态建模和多扩展目标的跟踪估计.首先, 结合广义标签多伯努利滤波器(Generalized labelled multi-Bernoulli, GLMB)建立了扩展目标的量测有限混合模型(Finite mixture models, FMM), 利用Gibbs采样和贝叶斯信息准则(Bayesian information criterion, BIC)准则推导出有限混合模型的参数来对多扩展目标形状进行学习, 然后采用等效量测方法来替代扩展目标产生的量测, 对扩展目标形状采用椭圆逼近建模, 实现扩展目标形状与状态的估计.仿真实验表明本文所给的方法能够有效跟踪多扩展目标, 并且在目标个数估计方面优于CBMeMBer算法.此外, 与标签多伯努利滤波(LMB)计算比较表明: GLMB和LMB算法滤波估计精度接近, 二者精度高于CBMeMBer算法.     

Rao-Blackwellized粒子势均衡多目标多伯努利滤波器

由于多伯努利滤波器直接近似递推了多目标状态的后验概率密度,使得多目标跟踪问题在基于随机有限集理论框架下的求解及目标状态的估计显得更为直观.本文针对一个状态可分解(线性/非线性)的状态空间模型,分析基于Rao-Blackwell定理的滤波估计方法,结合噪声的去相关构造线性状态的滤波方程.文中详细推导并提出Rao-Blackwellized粒子势均衡多目标多伯努利滤波器的一般实现形式,包括给出多伯努利非线性状态粒子滤波的实现形式,并结合非线性滤波结果给出多伯努利线性状态的递推滤波公式.本文提出的滤波器实现方法能够在更低维的状态空间上进行采样,滤波器的整体跟踪性能得到提高.多目标跟踪的仿真实验结果验证了该算法的有效性.     

基于变分贝叶斯势均衡多目标多伯努利滤波的 多扩展目标跟踪算法

由于传统多扩展目标跟踪算法在量测噪声协方差未知情况下跟踪性能急剧下降,本文提出一种基于变分贝叶斯的势均衡多目标多伯努利滤波(VB--CBMe MBer)跟踪算法,并给出了高斯混合实现.该算法在未知量测噪声协方差的情况下,将量测看成随机分布在扩展目标上的量测产生点所产生,利用变分贝叶斯方法近似地求出各量测产生点状态和量测噪声协方差的联合概率密度,并给出其递归形式以估计量测产生点,继而将得到的量测产生点状态进行聚类得到扩展目标的状态.仿真实验表明,所提算法可以自适应地跟踪未知数目、未知量测噪声协方差的多扩展目标.其跟踪精度与传统的CBMe MBer跟踪算法相比,有明显提高.     

面向非线性MTT的多模型泊松多伯努利混合滤波算法

在多目标跟踪(Multi-target Tracking, MTT)的非线性特性与低检测概率情况下,针对多伯努利滤波算法的高斯混合(Gaussian Mixture, GM)实现难以精确估计目标的势与运动状态的实际问题,本文提出了一种适用于非线性系统的泊松多伯努利混合滤波(Poisson Multi-Bernoulli Mixture Filter, PMBM)算法.首先,推导出多模型泊松多伯努利混合滤波的高斯混合(GM Multi-model PMBM,GM-MM-PMBM)实现过程.然后,分别对GM-MM-PMBM的伯努利高斯分量进行预测与更新,实现了基于非线性系统的MTT.为提升系统稳定性,基于平方根协方差矩阵推导出GM-MM-PMBM均方根容积卡尔曼滤波算法的实现过程.最后,仿真实验综合验证了本文算法的跟踪性能.     

多目标跟踪中多传感器分布式控制策略

分布式传感器网络技术在复杂的多目标跟踪系统中发挥了极其重要的作用.本文针对多传感器多目标跟踪中的分布式传感器控制问题,提出了基于信息论的多传感器控制策略.首先,本文利用随机有限集(RFS)建模,给出了多传感器多伯努利滤波器,并通过一组参数化的多伯努利过程来近似多传感器多伯努利密度.进一步的,通过多伯努利滤波器的序贯蒙特卡罗实现,设计采样方案对多伯努利密度进行粒子采样,用一组带有权值的粒子集近似多目标状态空间分布.随后,借助巴氏距离(Bhattacharyya distance)作为传感器控制的评价函数用于多个传感器的独立并行控制方案的决策制定.作为另外一个重要内容,本文提出了一种基于多目标战术重要性评估的多传感器控制策略.该控制方案旨在评估多目标战术重要性的基础上对威胁度最大的目标进行优先跟踪.最后,仿真实验验证了所提算法的有效性.     

基于随机有限集的多扩展目标跟踪研究进展

对基于随机有限集的多扩展目标跟踪方法的研究现状和进展进行综述.首先给出扩展目标的数学模型;然后给出扩展目标形状估计的3种方法:随机矩阵法,随机超平面法和高斯过程法;接着给出多扩展目标的随机有限集滤波器的算法及优缺点;最后介绍多扩展目标跟踪的主要应用,并对其未来的发展方向做进一步展望.     

基于GGIW-PMB的衍生扩展目标跟踪

针对标准的扩展目标泊松多伯努利(Poisson multi-Bernoulli, PMB)滤波器难以有效跟踪衍生目标的问题,提出一种改进的PMB跟踪算法.算法采用随机矩阵法对扩展目标外形和尺寸建模,在滤波预测阶段利用多假设模型对衍生事件进行预测,得到多个伽玛高斯逆威沙特(gamma Gaussian inverse Wishart, GGIW)预测假设分量,最后在滤波更新阶段对预测分量更新得到扩展目标的运动状态和扩展形状估计.仿真结果表明,与标准的PMB滤波算法相比,所提算法有效改善衍生扩展目标的跟踪性能.     

机动多目标跟踪中的传感器控制策略的研究

针对机动多目标跟踪中的传感器控制问题, 本文提出一种基于信息论的多模型多伯努利滤波器的控制方案. 首先, 基于随机有限集(Random finite set, RFS)方法给出信息论下的传感器控制的一般方法; 其次, 本文给出多模型势均衡多目标多伯努利滤波器的序贯蒙特卡罗实现形式. 此外, 提出一种目标导向的多伯努利概率密度的粒子采样方法, 并借助该方法近似多目标概率密度, 继而利用Bhattacharyya 距离求解最终的控制方案. 典型机动多目标跟踪问题的仿真应用验证了本文传感器控制方法的有效性.     

概率逻辑学基本定理的推广及其应用

利用赋值集的随机化方法,在[n]值命题逻辑系统中提出了公式的随机不可靠度概念,证明了一个有效推理结论的随机不可靠度不大于各前提的随机不可靠度与其必要度的乘积之和。通过推广的概率逻辑学基本定理,证明了随机逻辑度量空间中MP规则和HS规则成立。     

C语言中互不相同随机数的应用研究

文章对随机数的应用问题进行了详细的分析,给出了一种实现的算法,并用C语言实现。通过该问题的C实现,可使学习者清晰地观测到解决该问题的全过程。     

基于Markov过程预测的页面置换方法

页面替换是操作系统实现虚拟存换方法的缺点,提出一种基于Markov随机过程的页面调度模型.通过分析页面置换序列,发现其规律并模拟该序列,目的是接近最优替换方法以降低缺页率.实现了基于该模型的页面置换方法,通过实验证明新方法能提高页面换入命中率并控制时间耗费.     

随机谕言模型

介绍了可证明安全理论中的重要工具——随机谕言模型,包括随机谕言模型的起源、基本性质和方法、随机谕言模型中的代表方案、明文知晓性质、随机谕言的实例化、随机谕言不可实例化的性质和相关负面结论以及对弱化的随机谕言模型的研究.此外,比较了随机谕言模型和其他理想模型,简介了标准模型中的方案设计状况.     

随机约束满足问题相变研究综述

随机约束满足问题是经典的NP完全问题,在理论研究和现实生活中有着广泛应用。研究人员发现随机约束满足问题存在相变现象,近几十年来关于此问题相变的研究成果不断涌现。从随机图着色问题和随机可满足问题2个最经典的随机约束满足问题入手,从算法研究、理论物理和数学证明3个方面综述了随机图着色问题和随机可满足问题的相变研究成果。最后对随机约束满足问题相变的研究趋势进行了展望。     

Morphological analysis of discrete random shapes

In this paper I discuss a number of theoretical issues regarding the morphological analysis of discrete random shapes by means of Matheron's random set theory. I revisit this theory by limiting myself to the discrete case, since most image data are available in a discrete form. Although it may seem that the transition from the continuous to the discrete case is straightforward (since most of Matheron's theory is general enough to incorporate the discrete case as a special case), this transition is often challenging and full of exciting and, surprisingly, pleasant results. I introduce the concept of the cumulative-distribution functional of a discrete random set and review some fundamental properties of the capacity functional (a fundamental statistical quantity that uniquely defines a random set and relates random set theory to mathematical morphology). In analogy to a recent result and under a natural boundness condition, I show that there exists a one-to-one correspondence between the probability-mass function of a discrete binary random field and the corresponding cumulative-distribution functional. The relationship between the cumulative-distribution functional and the capacity functional of a discrete random set is also established. The cumulative-distribution and capacity functionals are related to the higher-order moments of a discrete binary random field, and, therefore, their computation is equivalent to computing these moments. A brief discussion of how to perform such computations for a certain class of discrete random sets is provided. The capacity functional of a morphologically transformed, continuous random set cannot be associated to the capacity functional of the random set itself, except in the case of dilation. I show that the derivation of such an association is possible in the discrete case and for the cases of dilation and erosion and more complicated morphological transformations, such as opening and closing. These relationships are then used to derive a fundamental result regarding the statistical behavior of opening and closing morphological filters. I also show that the probability-mass function of a discrete binary random field may be expressed in terms of the cumulative-distribution functional or the capacity functional of a morphologically transformed discrete random set by means of a hit-or-miss transformation. I also introduce moments for discrete random sets, which permit generalization of the concepts of autocorrelation and contact distribution. Furthermore, I demonstrate the fact that the class of opening-based size distributions, introduced axiomatically by Matheron, are higher-order moments of a discrete random set, therefore statistically demonstrating that size distributions are good statistical summaries for shape. Finally, convex random sets are viewed in the discrete domain. My final result regarding convexity, is similar to Matheron's. However, the tools used here for the derivation of such a result are different from the ones used by Matheron, whose approach to this subject is limited to the continuous case.     

A probabilistic scheme of independent random elements distributed over a finite lattice. I. Exact probability distributions of functionals of union of random elements

A probabilistic scheme of independent random elements with values in a finite lattice is introduced. For the scheme, exact expressions of probability distributions of a functional of general form and union of random corank elements are obtained. Various probabilistic combinatorial problems (on probability distributions of the number of uncovered points in a generalized scheme of grouped distribution of particles, on the number of connectivity components of a random hypergraph, on the number of solutions of a system of random linear equations over a finite ring with unity, etc.) are demonstrated to be naturally formulated in terms of the scheme constructed.Translated from Kibernetika i Sistemnyi Analiz, No. 5, pp. 3–15, September–October 2004.     

VFP数据库加密算法的研究与实现

对于最初给定的随机种子,随机函数会生成一个相应的数列。数据库数据随机加密,是一种利用随机函数的随机数序列与数据库数据进行异或等运算,实现数据库数据加密的方法。本文对数据库数据随机加密算法进行了研究和探讨,给出了在VFP中实现数据库数据随机加密、解密的程序代码。     

Orthogonal decompositions of 2-D nonhomogeneous discrete random fields

Imposing atotal-order on a two-dimensional (2-D) discrete random field induces an orthogonal decomposition of the random field into two components: Apurely-indeterministic field and adeterministic one. The purely-indeterministic component is shown to have a 2-D white-innovations driven moving-average representation. The 2-D deterministic random field can be perfectly predicted from the field's past with respect to the imposed total-order definition. The deterministic field is further orthogonally decomposed into anevanescent field, and aremote past field. The evanescent field is generated by the columnto-column innovations of the deterministic field with respect to the imposed nonsymmetrical-half-plane total-ordering definition. The presented decomposition can be obtained with respect to any nonsymmetrical-half-plane total-ordering definition, for which the nonsymmetrical-half-plane boundary line has rational slope.