基于DEnKF的背景误差协方差局地化和协方差膨胀研究

尽管DEnKF同化不会引入观测采样误差,但小集合仍会造成背景误差协方差矩阵存在伪相关,出现滤波发散。为了减少小集合对数据同化结果的影响,结合Lorenz96模型和DEnKF同化方案分析了协方差局地化和协方差膨胀方法对背景误差协方差矩阵、增益矩阵及同化结果的影响。实验表明:协方差局地化方法能消除背景误差协方差矩阵和增益矩阵中的伪相关,增大背景误差协方差矩阵的秩,有助于滤波算法收敛到真实解;而协方差膨胀方法不能消除背景误差协方差矩阵和增益矩阵中的伪相关,只能改善在每个同化周期内背景误差协方差系统性被低估的现象;同化过程中采用合适的局地化半径和方差膨胀因子能够较好地改善同化结果的精度。     

一种高精度APNCKF算法在谐波检测中的应用

为了进一步提高含噪环境下谐波检测的精确度,提高卡尔曼滤波器的稳定性,对系统噪声协方差进行了分析,通过不断的在线辨识出过程噪声协方差,提出了一种自适应过程噪声协方差卡尔曼滤波算法。该算法利用序贯最大化可信度更新先验信息来辨识过程噪声,然后通过卡尔曼滤波器进行迭代运算,估计出相应的幅值和相位。该算法最大的特点就是辨识出的过程噪声Q的骤然增大匹配的即是谐波幅值暂降的出现。通过在MATLAB环境下进行谐波仿真验证,结果表明该算法在准稳态条件下较好地跟踪电力系统谐波状态,且与常规卡尔曼、基于最大似然准则的卡尔曼、小波/小波包变换相比,该自适应算法的收敛速度较快、滤波精度高、实时性以及稳定性较好,具有重要的工程实际意义。     

Accuracy analysis of a covariance matching approach for identifying errors-in-variables systems

A system identification method for errors-in-variables problems based on covariance matching was recently proposed. In the first step, a small amount of covariances of noisy input–output data are computed, and then a parametric model is fitted to these covariances. In this paper, the method is further analyzed and the asymptotic accuracy of the parameter estimates is derived. An explicit algorithm for computing the asymptotic covariance matrix of the parameter estimates is given, and the identification method is shown to be asymptotically statistically efficient assuming that the given information is the computed covariances. As an important byproduct, an efficient algorithm is presented for computing the covariance matrix of the computed input–output covariances.     

Efficient covariance matrix update for variable metric evolution strategies

Randomized direct search algorithms for continuous domains, such as evolution strategies, are basic tools in machine learning. They are especially needed when the gradient of an objective function (e.g., loss, energy, or reward function) cannot be computed or estimated efficiently. Application areas include supervised and reinforcement learning as well as model selection. These randomized search strategies often rely on normally distributed additive variations of candidate solutions. In order to efficiently search in non-separable and ill-conditioned landscapes the covariance matrix of the normal distribution must be adapted, amounting to a variable metric method. Consequently, covariance matrix adaptation (CMA) is considered state-of-the-art in evolution strategies. In order to sample the normal distribution, the adapted covariance matrix needs to be decomposed, requiring in general Θ(n ³) operations, where n is the search space dimension. We propose a new update mechanism which can replace a rank-one covariance matrix update and the computationally expensive decomposition of the covariance matrix. The newly developed update rule reduces the computational complexity of the rank-one covariance matrix adaptation to Θ(n ²) without resorting to outdated distributions. We derive new versions of the elitist covariance matrix adaptation evolution strategy (CMA-ES) and the multi-objective CMA-ES. These algorithms are equivalent to the original procedures except that the update step for the variable metric distribution scales better in the problem dimension. We also introduce a simplified variant of the non-elitist CMA-ES with the incremental covariance matrix update and investigate its performance. Apart from the reduced time-complexity of the distribution update, the algebraic computations involved in all new algorithms are simpler compared to the original versions. The new update rule improves the performance of the CMA-ES for large scale machine learning problems in which the objective function can be evaluated fast.     

基于指纹库更新补偿的波束跟踪算法研究

针对拒止、复杂电磁环境下,高动态无人节点定向通信面临的坐标信息不精确、飞行姿态和轨迹变化剧烈等问题,为保持可靠的波束对准与跟踪,提出了一种基于卡尔曼滤波的指纹库更新补偿算法。首先,利用卡尔曼滤波算法对自身姿态进行预测更新,建立新的载体坐标系;利用改进的算法对波束指向进行预测更新,并利用指纹库对状态向量均值和协方差矩阵进行更新补偿,调节采样比例,并将新的数据存入指纹库对指纹库数据更新,然后进行二次状态信息预测,完成最终波束指向。整体设计的波束跟踪算法流程更加符合实际应用场景,满足无人机自组网的需求。仿真结果表明,在半波束宽度为3°,100个通信时隙中,维持正常通信的成功率有92%以上,相比传统跟踪算法提高了8%,具有更加稳定的通信质量。     

Event based distributed kalman filter for limited resource multirobot cooperative localization

This paper presents a multirobot cooperative event based localization scheme with improved bandwidth usage in a heterogeneous group of mobile robots. The proposed method relies on an agent based framework that defines the communications between robots and on an event based Extended Kalman Filter that performs the cooperative sensor fusion from local, global and relative sources. The event is generated when the pose error covariance exceeds a predefined limit. By this, the robots update the pose using the relative information available only when necessary, using less bandwidth and computational resources when compared to the time based methods, allowing bandwidth allocation for other tasks while extending battery life. The method is tested using a simulation platform developed in the programming language JAVA with a group of differential mobile robots represented by an agent in a JADE framework. The pose estimation performance, error covariance and number of messages exchanged in the communication are measured and used to compare the traditional time based approach with the proposed event based algorithm. Also, the compromise between the accuracy of the localization method and the bandwidth usage is analyzed for different event limits. A final experimental test with two SUMMIT XL robots is shown to validate the simulation results.     

基于信任机制的水下传感器网络节点安全定位算法

为了及时检测出水下传感器网络(UWSN)定位系统中的恶意锚节点,提出一种基于信任机制的节点安全定位算法。算法结合簇结构和信任机制,根据锚节点提供的位置信息采用Beta分布作出初步信任评价,并可根据需要调整信任更新权重。为了降低了水声信道的不稳定性对信任评价过程的影响,同时识别恶意锚节点的信任欺骗行为,提出信任过滤机制 (TFM),对直接信任值进行差异量化,由簇头节点决定各锚节点是否可信。仿真结果表明所提算法适用于水下传感器网络,并且能够及时识别恶意锚节点,在定位系统的精确度和安全性方面都有很大提升。     

基于协方差矩阵的盲分离算法

提出了一种新的实时线性混叠信号的盲分离算法，该算法利用信号相互独立时其协方差矩阵的对角化特征作为分离准则，采用最速下降法进行分离。该算法对源信号和混叠矩阵没有过多要求且计算量不大，理论分析与仿真结果表明，该算法具有很好的分离效果。     

Estimation of the conditional risk in classification: The swapping method

The bias of the empirical error rate in supervised classification is studied. It is shown that this bias can be understood as a covariance between the classification rule and the labeling of the training data. From this result, a new penalized criterion is proposed to perform model selection in classification. Applications of the resulting algorithm to simulated and real data are presented.     

无线传感器网络中基于虚拟力的分布式节点定位

节点定位是无线传感器网络应用中需要解决的一个基本问题。传统算法大都基于集中式方法估计节点位置,从而导致较大开销。因此,结合最小二乘法进行初步估计定位,并在此基础上,给出了基于虚拟力的传感器节点定位模型,提出了基于虚拟力的分布式定位算法,该算法通过邻居节点间信息的分布式交互,能够有效节省定位开销。进一步,在定位过程中引入未知节点升级机制,以提高收敛速度。一系列仿真实验表明,该算法能够通过分布式迭代定位,快速实现全网节点的精确定位。     

Improved Stein-type shrinkage estimators for the high-dimensional multivariate normal covariance matrix

Many applications require an estimate for the covariance matrix that is non-singular and well-conditioned. As the dimensionality increases, the sample covariance matrix becomes ill-conditioned or even singular. A common approach to estimating the covariance matrix when the dimensionality is large is that of Stein-type shrinkage estimation. A convex combination of the sample covariance matrix and a well-conditioned target matrix is used to estimate the covariance matrix. Recent work in the literature has shown that an optimal combination exists under mean-squared loss, however it must be estimated from the data. In this paper, we introduce a new set of estimators for the optimal convex combination for three commonly used target matrices. A simulation study shows an improvement over those in the literature in cases of extreme high-dimensionality of the data. A data analysis shows the estimators are effective in a discriminant and classification analysis.     

伪多跳中继分布式一致性算法

为了提高分布式一致性问题的收敛速度, 减少节点间的通信成本, 本文提出一种分布式一致性算法. 新算法采用单跳通信, 利用非邻接节点的前状态信息进行节点状态更新. 本文证明了在无向通信拓扑下新算法的一致性收敛, 并分析计算了新算法的通信量以及收敛速度的变化. 通过理论分析和仿真验证, 结果表明新算法具有收敛速度快, 通信方式简单, 通信量少, 存在通信延时情况下通信数据不容易丢失等优点.     

协作多点传输中基于空间协方差动态更新的码本自适应方法

针对典型码本自适应算法的信道协方差矩阵反馈时间间隔过长,导致高速环境下系统性能迅速恶化的缺点,提出了一种基于信道协方差矩阵动态更新的码本自适应改进算法。基于协方差矩阵和码本之间的等效关系和均方误差最小的原则,在接收端将协方差矩阵拆分成一个固定参数和一个码本向量,并向发射端反馈该码本向量。然后,在发射端利用反馈的码本向量及本地固定参数和上一时刻的协方差矩阵,重构当前时刻的协方差矩阵。仿真结果表明,较之典型算法,提出的算法具有更好的性能表现,尤其在高速环境下,可以获得近2dB的增益。     

Energy-efficient localization and target tracking via underwater mobile sensor networks

Underwater mobile sensor networks (UMSNs) with free-floating sensors are more suitable for understanding the immense underwater environment. Target tracking, whose performance depends on sensor localization accuracy, is one of the broad applications of UMSNs. However, in UMSNs, sensors move with environmental forces, so their positions change continuously, which poses a challenge on the accuracy of sensor localization and target tracking. We propose a high-accuracy localization with mobility prediction (HLMP) algorithm to acquire relatively accurate sensor location estimates. The HLMP algorithm exploits sensor mobility characteristics and the multi-step Levinson-Durbin algorithm to predict future positions. Furthermore, we present a simultaneous localization and target tracking (SLAT) algorithm to update sensor locations based on measurements during the process of target tracking. Simulation results demonstrate that the HLMP algorithm can improve localization accuracy significantly with low energy consumption and that the SLAT algorithm can further decrease the sensor localization error. In addition, results prove that a better localization accuracy will synchronously improve the target tracking performance.     

基于增量式特征基背景模型的运动目标检测

背景差法是目标运动检测的主流方法,关键在于背景模型自适应更新.针对传统特征基背景模型批处理方式计算量大、更新速度慢的问题,采用增量式主成分分析来建立特征基背景模型.首先计算样本图像的初始背景图像,然后采用CCFIPCA算法更新特征基背景模型,最后通过输入帧和重建帧的欧氏距离检测前景运动目标.算法以视频帧整体来建立背景模型,克服了混合高斯模型和核密度估计以孤立像素点建模的不足,提高了背景建模的鲁棒性.在SIMULINK下的仿真实验表明,算法能很好地适应高速公路交通场景动态变化,在有光线变化和阴影影响的情况下能完整、准确地提取出运动车辆轮廓.     

Controllability and observability covariance matrices for the analysis and order reduction of stable nonlinear systems

This paper presents a framework for nonlinear systems analysis that is based upon controllability and observability covariance matrices. These matrices are introduced in the paper and it is shown that gramians for linear systems form special cases of the covariance matrices. The covariance matrices can be transformed via a balancing-like transformation and nonlinearity measures are defined based upon these transformed covariance matrices. Subsequently, the covariance matrices are used for reduction of the nonlinear model. It is shown that the model reduction procedure reduces to balanced model truncation for linear systems for impulse inputs. Furthermore, it is also shown that several model reduction procedures that were developed by other researchers, and assumed to be independent from one another, are related. The findings are illustrated with an example.     

Nonparametric Stein-type shrinkage covariance matrix estimators in high-dimensional settings

Estimating a covariance matrix is an important task in applications where the number of variables is larger than the number of observations. Shrinkage approaches for estimating a high-dimensional covariance matrix are often employed to circumvent the limitations of the sample covariance matrix. A new family of nonparametric Stein-type shrinkage covariance estimators is proposed whose members are written as a convex linear combination of the sample covariance matrix and of a predefined invertible target matrix. Under the Frobenius norm criterion, the optimal shrinkage intensity that defines the best convex linear combination depends on the unobserved covariance matrix and it must be estimated from the data. A simple but effective estimation process that produces nonparametric and consistent estimators of the optimal shrinkage intensity for three popular target matrices is introduced. In simulations, the proposed Stein-type shrinkage covariance matrix estimator based on a scaled identity matrix appeared to be up to 80% more efficient than existing ones in extreme high-dimensional settings. A colon cancer dataset was analyzed to demonstrate the utility of the proposed estimators. A rule of thumb for adhoc selection among the three commonly used target matrices is recommended.     

基于RSSI测距误差修正的改进型DV-distance差分定位算法

DV-distance是一种基于多跳机制的定位算法,其相邻节点间的距离通过RSSI测距技术实际测量得到。为了减少RSSI测距误差对定位精度的影响,首先对RSSI测距误差进行修正,再对已有的信标节点间计算距离误差修正值的方法作进一步改进,提出一种基于RSSI测距误差修正的改进型DV-distance差分定位算法。仿真结果表明,相比已有的定位算法,该改进算法不仅能提高节点的定位精度,还能减少网络通信开销及计算复杂度。     

考虑通信延迟的多自治水下航行器协同定位算法

自治水下航行器(AUV)协同定位中通信延迟具有常态性. 面对延迟到达的信息, 传统方法一般会有定位精 度或实时性的损失. 针对通信延迟的不利影响, 本文在建立水声探测和通信时延模型的基础上, 以扩展卡尔曼滤波 (EKF)为算法框架, 提出了信息顺序到达和信息出序到达2种协同定位算法, 并以建构面向信息出序情景的算法为 主要创新工作. 在信息顺序到达算法中, 将延迟信息进行序贯处理以减小定位误差. 在信息出序到达算法中, 以信 息出现一步滞后的延迟为背景, 使用出序信息直接对从AUV最新状态估计进行再更新, 信息无损地实时估计运动 状态. 计算机仿真实验结果表明, 本文算法相比于传统的航位推算、整周期滤波、量测丢弃等方法, 具有更高的估计 精度; 相比于数据缓存滤波、重新滤波等方法, 具有强实时性.     

高斯过程下的CMA-ES在医学图像配准中的应用

为了改进协方差矩阵自适应进化策略(CMA-ES)的性能,提出了一种高斯过程协助下的协方差矩阵自适应进化策略(GPACMA-ES)。该策略利用CMA-ES中的协方差矩阵构建核函数,引入高斯过程,在线学习历史经验,并根据历史经验预测全局最优解的最有前景区域,有效地降低了适应度函数的评价次数。同时,为了提高群体的搜索效率,引入了置信区间。群体在置信区间内更高效地采样,使得算法具备更快的收敛速度和全局寻优能力。最后,将GPACMA-ES算法应用于医学图像配准中,配准精度和效率均高于标准的CMA-ES算法。