基于D-S证据理论的特征融合车标识别方法

在针对现有的智能交通对车辆多维信息识别存在识别精度不高的情况,特别是对于车标识别,很大程度上识别结果依赖于高分辨和高质量的图像.提出了一种新的车标识别方法,用于识别卡口捕获的低质量车标图像,该方法是基于D-S证据理论的特征融合方法,提取Hu不变矩和HOG特征,采用不同的分类器构造基本概率分配（BPA）,采用改进D-S证据理论进行融合,根据判别规则给出最终的识别结果.通过实验证明在低分辨的情况下仍能保持较高的准确率,分类准确率达94.29%,相比单一的特征识别,具有更强的鲁棒性.     

Intuitionistic fuzzy evidential power aggregation operator and its application in multiple criteria decision-making

The theory of intuitionistic fuzzy sets (IFS) is widely used for dealing with vagueness and the Dempster--Shafer (D-S) evidence theory has a widespread use in multiple criteria decision-making problems under uncertain situation. However, there are many methods to aggregate intuitionistic fuzzy numbers (IFNs), but the aggregation operator to fuse basic probability assignment (BPA) is rare. Power average (P-A) operator, as a powerful operator, is useful and important in information fusion. Motivated by the idea of P-A power, in this paper, a new operator based on the IFS and D-S evidence theory is proposed, which is named as intuitionistic fuzzy evidential power average (IFEPA) aggregation operator. First, an IFN is converted into a BPA, and the uncertainty is measured in D-S evidence theory. Second, the difference between BPAs is measured by Jousselme distance and a satisfying support function is proposed to get the support degree between each other effectively. Then the IFEPA operator is used for aggregating the original IFN and make a more reasonable decision. The proposed method is objective and reasonable because it is completely driven by data once some parameters are required. At the same time, it is novel and interesting. Finally, an application of developed models to the ‘One Belt, One road’ investment decision-making problems is presented to illustrate the effectiveness and feasibility of the proposed operator.     

基于人工免疫与D-S证据理论的故障诊断方法

针对使用信息融合技术进行故障诊断时,基本概率赋值难以确定的问题,提出一种基于人工免疫原理与D-S证据理论相结合的故障诊断方法.该方法使用无类标数据进行人工免疫聚类,构建每一个传感器的故障空间的分类模型,设计了中心抗体和一般抗体的识别半径对未知样本进行免疫识别,计算该样本在各故障分类器下的基本概率赋值,最后通过D-S证据理论将各基本故障概率赋值进行融合诊断,基于信任函数进行故障决策.试验结果表明该方法可以处理多信息源数据,提高了故障识别能力,有一定实践意义.     

一种有效解决D-S理论冲突证据合成的方法

针对标准D-S证据理论中存在的冲突证据合成问题,提出了一种冲突证据加权的方法。该方法将各个传感器的基本概率赋值映射到多维空间中的某个点,计算任意两点的欧氏距离,利用平均距离确定证据的权重,采用哈夫曼树对证据加权平均后再利用D-S合成规则实现信息的融合。实例论证了该方法的有效性,它能有效解决冲突证据合成的问题。     

Attribute measure recognition approach and its applications to emitter recognition

1Introduction Radar emitter recognition has become an important issue in military intelligence,surveillance,and reconnaissance.With the rapid development of radar technology,the density and complexity of radar signal are increasing.Moreover,radar signals take on uncertainty,illegibility and contradiction.Current algorithms for radar emitter recogni-tion do not always give good performance.So some researches have been conducted for emitter recognition over the past years,such as expert system,…     

基于SVM概率输出与证据理论的多分类方法

单一技术无法有效解决多类分类问题。为此,提出一种基于一对多支持向量机(SVM)的基本概率分配输出方法,并与置信最大熵模型的D-S证据组合方法结合,给出基于SVM概率输出和证据理论的多分类模型。在3种UCI标准数据集上的仿真结果表明,该方法的分类精度优于传统的一对多和一对一硬输出方法,是一种有效的多类分类方法。     

Quantum model of mass function

Dempster-Shafer (D-S) evidence theory has been used in many fields due to the flexibility and effectiveness in modelling uncertainties, which is the extension of classical probability. Uncertainty principle is one of the most important principles in quantum theory, which has been used in many fields. How to set the connection between quantum theory and D-S evidence theory is also an open issue. Hence, the paper proposed the quantum model of mass function to consider the quantum theory and D-S evidence theory. In the proposed quantum method, quantum mass function uses euler formula to represent. The paper also discusses some operations based on the quantum model of the mass function. Moreover, the paper also discusses the relationship between quantum mass function and classical mass function by using some numerical examples. Classical mass function is a special case when there is no interference in quantum mass function. Similar to the other quantum models, this study provides a more wide application in quantum information.     

基于情感特征分类的语音情感识别研究

针对语音信号的实时性和不确定性,提出证据信任度信息熵和动态先验权重的方法,对传统D-S证据理论的基本概率分配函数进行改进;针对情感特征在语音情感识别中对不同的情感状态具有不同的识别效果,提出对语音情感特征进行分类。利用各类情感特征的识别结果,应用改进的D-S证据理论进行决策级数据融合,实现基于多类情感特征的语音情感识别,以达到细粒度的语音情感识别。最后通过算例验证了改进算法的迅速收敛和抗干扰性,对比实验结果证明了分类情感特征语音情感识别方法的有效性和稳定性。     

D-S证据理论在审计证据融合中的应用研究

古典概率难以解释审计判断的不确定性,而D-S证据理论是进行不确定性推理的有效方法,因此应用D-S证据理论进行审计证据融合的研究。针对审计证据的组合问题,提出了基于三角形模糊隶属度函数的基本概率分配函数计算方法,给出了证据组合结果的判决规则,并通过实例验证了该方法的有效性。     

An improved soft likelihood function for Dempster–Shafer belief structures

Information fusion is an important research direction. In the field of information fusion, there are many methods for evidence combination. Recently, Yager proposed a method of soft likelihood function to combine probabilistic evidence effectively. Considering that basic probability assignment (BPA) can deal with uncertainty information more effectively, in this paper, we extend Yager's soft likelihood function to combine BPA. First, according to the BPA evaluations of evidence sources, belief function and plausibility function on each alternative are calculated. Then, interval numbers are constructed by the obtained belief function and plausibility function to indicate the belief interval on each alternative. Next, the descending sorting of interval numbers is aggregated by the ordered weighted averaging operator. Finally, by sorting the result of the aggregation, the ordering of alternatives is obtained. A numerical example and an example of application in Iris data set classification illustrate the effectiveness of the improved method.     

改进证据分类合成方法在农作物生长环境评估中的应用

针对传统DS证据理论存在处理冲突证据的不足,基于证据间的相似度引入了信 息熵属性,修正了证据分类属性,结合证据间相似度属性将证据集重新划分为可信度高证据 、一般性证据和冲突证据,对分类的证据集赋予不同的重要性系数,并加以修正改进。改进后使得一般性证据和高冲突证据向可信度高的证据意见靠拢,最后利用DS组合规则对于修正后的证据进行合成。针对农作物生长环境中多个传感器获取的数据构造其所对应证据的基本概率分配函数,利用模糊理论对基本概率分配函数进行取值。实验采用各类传感器测得的真实数据集进行实验,结果表明改进的方法既能够很好地解决冲突问题,同时能降 低证据的不确定性。     

A new probability transformation method based on a correlation coefficient of belief functions

The Dempster-Shafer evidence theory is widely used in many fields of information fusion because of its advantage in handling uncertain information. One of the key issues in this theory is how to make decision based on a basic probability assignment (BPA). Currently, a feasible scheme is transforming a BPA to a distribution of probabilities. However, little attention was paid to the correlation between BPA and probability distribution. In this paper, a novel method about the probability transformation based on a correlation coefficient of belief functions is proposed. The correlation coefficient is a new measurement, which can effectively measure the correlation between BPAs. The proposed method aims at maximizing the correlation coefficient between the given BPA and the transformed probability distribution. On the basis of this idea, the corresponding probability distribution can be obtained and could reflect the original information of the given BPA to the maximum extent. It is valid to consider that the proposed probability transformation method is reasonable and effective. Numerical examples are given to show the effectiveness of the proposed method.     

Evaluation method based on fuzzy relations between Dempster–Shafer belief structure

Many relations in the real world can be described by mathematical language. Fuzzy set theory can transform human language into mathematical language and use membership degree function to describe relations between events. Dempster–Shafer evidence theory provides basic probability assignment (BPA), which can describe the occurrence rate of attributes in basic events. Based on the known membership degree function and BPA distribution, a new evaluation method is proposed in this paper to analyze decision making. Given the relations among relevant events, which are expressed by BPA distribution and membership degree function, the relations among basic events and top event can be obtained. The Dempster's combination rule and pignistic probability transformation are used to transform BPA distribution into probability distribution. The belief measure is applied to deal with these fuzzy relations. Some numerical examples are given in this paper to illustrate the proposed evaluation methodology.     

一种基于信息融合的多源遥感图像分类方法

基于D-S证据理论提出了一种多源遥感图像分类融合的新方法。首先通过人为选择感兴趣的分类区域，提取特征获取基本概率分配函数，将待分类的多源图像进行分类融合，从而得到最终的分类结果。试验表明，相比于K-mean分类方法，这种分类融合方法可以有效地减少分类过程中的不确定性信息，提高分类精度。     

D-S证据理论在目标身份识别中的应用

针对多源信息融合中目标身份属性识别问题,简要介绍了D-S证据理论框架,阐述了基于基本概率指派(BPAF)决策的目标身份属性融合策略、步骤,利用D-S合成规则得到融合后的基本概率指派,实现了多传感器信息融合.仿真实验证明了方法的有效性.     

基于ELM和证据理论的纹理图像分类

目前纹理图像分类有不同的方法,但对纹理的描述还不够全面,而且当有新方法提取的特征加入时,系统的可扩展性也不够,通用性不好。本文针对上述问题提出了一种将D-S证据理论与极限学习机相结合的决策级融合模型,用来对纹理图像进行分类。采用三种不同方法来提取特征以获得更多更全面的纹理表现形式,并对提取的每种特征向量用极限学习机建立相应的分类器,最后用D-S证据理论在不确定性表示、度量和组合方面有着的优势来进行决策级融合。对于证据理论中基本概率赋值函数(BPAF)难以有效获取的问题,由于极限学习机具有学习速度快,泛化性能好的优点并且产生唯一的最优解的优点,所以利用其来构造其基本概率赋值函数。实验结果表明这种方法比单个分类器具有更高的识别正确率,降低了识别的不确定性。     

超球SVM与证据理论在故障诊断中的应用

针对使用多传感器信息融合技术进行故障诊断时,故障模式较多、基本概率赋值难以确定的问题,提出一种基于超球支持向量机与D-S证据理论相结合的故障诊断方法。该方法使用超球支持向量机针对每一个传感器的故障空间训练分类模型,根据类内隶属度与类-类相似度得到各故障类别的基本概率赋值,利用D-S证据理论进行证据融合,基于信任函数进行故障决策。试验结果表明该方法提高了故障识别能力,有一定实践意义。     

考虑时间参数的网络安全态势评估模型研究

针对已有网络安全态势评估由于缺乏考虑数据源时变性而造成的证据源证据不可靠、评估出现误差的问题,引入时变函数来刻画多源证据的时变性,提出时变D-S证据理论,用以提高证据的可靠性。使用改进的时变D-S证据理论方法对多传感器的证据进行融合,得到威胁的发生概率,并在此基础上提出带有时间参数的网络安全态势评估层次化模型;最后利用网络实例数据,对所提出的网络安全态势评估模型进行了验证。实验对比结果表明,该模型对实际网络运行情况的评估更符合实际情况。     

多传感器数据融合技术在故障诊断中的应用

利用多传感器数据融合的方法进行故障诊断,建立融合故障诊断系统.将故障诊断系统按数据融合的方法分为数据级融合模块、特征级融合模块和决策级融合模块.数据级融合模块主要对多传感器的测量信号进行处理,提取出故障诊断的特征信息.特征级融合模块采用3个结构相同的并行神经网络,一是进行局部诊断;二是获得决策级D-S证据理论的基本概率赋值.决策级采用D-S证据理论的方法对特征级局部诊断的结果加以融合,得到最终的诊断结果.利用此系统在汽轮机转子试验台架上进行了故障诊断,得到了令人满意的结果.     

基于D-S证据理论的目标识别融合应用

基于D-S证据理论的目标识别融合系统,可以充分发挥多传感器信息的优势,提高目标识别结果的准确性.本文结合工程实践,分析地面目标融合识别过程中经典D-S证据理论方法处理数据出现的问题,发现使用D-S证据理论对于高冲突证据融合结果准确性较低.因此提出一种基于D-S证据理论的改进数据融合方法,将冲突因子与支持度标准偏差的相反数相乘,再与所有证据和乘积的正交相加,然后减去证据的基本概率的最大差.如果证据的冲突越大,这种方法的优势就越明显.如果证据中不存在冲突,则融合结果与原始D-S证据理论的项目一致.实验的比较数据表明,改进的信息融合方法对于改进解决冲突问题必不可少,并且是有效的.