快速对角阵权系数协方差交叉融合容积卡尔曼滤波器

针对互协方差信息未知的多传感器系统,本文提出了一种快速对角阵权系数协方差交叉融合算法(FDCI).本文首先提出了一种对角阵权系数协方差交叉融合(DCI)方案,并证明了所提出DCI算法在融合估计精度上高于经典批处理CI融合(BCI)算法.在此基础之上,针对非线性等复杂的互协方差未知的多传感器系统,提出FDCI算法,并证明了所提出FDCI算法的无偏性及鲁棒精度. FDCI融合算法虽然在融合估计精度上低于DCI,但FDCI无需进行多权系数的非线性代价函数的优化问题,进而大大降低了计算负担,提高了系统的实时性.最后,结合容积卡尔曼滤波算法(CKF)提出了快速对角阵权系数协方差交叉融合容积卡尔曼滤波算法.仿真实例验证了所提出算法的正确性和有效性.     

基于支持度和自适应加权的MEMS陀螺信息融合算法

针对MEMS陀螺阵列进行信息融合可以大幅提高其测量精度和可靠性,而传统信息融合算法大部分依赖于观测信息的先验知识而受到应用上的限制。在不依赖于先验知识的基于支持度的融合算法的基础上,提出一种新的加权系数构造方法,利用自适应加权法的思想改变各个传感器观测值方差对权系数的影响程度,既保证了阵列输出结果的可靠性,又能使融合后得到目标参数的总方差最小;而且自适应加权算法中,对各个传感器方差的计算考虑到环境因素等带来的噪声,与各传感器方差的真实值更加接近。实验结果表明,本文提出的融合算法优于传统的平均值估计融合算法和支持度融合算法。     

基于粒子群优化的大规模传感器网络节点调度策略

针对资源受限条件下大规模无线传感器网络中协作目标跟踪问题,提出一个基于粒子群优化的节点调度方案.该方案利用高斯粒子滤波算法和方差交叉融合算法获得目标状态预测信息,进而选择下一时刻簇成员节点,并构造了通信能耗的代价函数,利用粒子群优化方法选择最佳的簇头节点,减少了节点调度的计算复杂度,同时保持了较好的跟踪精度.仿真结果验证了所提出方案的有效性.     

多传感器信息融合Wiener滤波器和平滑器

应用现代时间序列分析方法,基于ARMA新息模型和增广状态空间模型,应用标量加权最优融合准则,对于带白色和有色观测噪声的ARMA信号,提出了多传感器分布式最优信息融合Wiener滤波器和平滑器,其中给出了计算局部平滑误差方差和互协方差的计算公式,它们可被用于计算最优加权系数。同单传感器情形相比,可提高平滑器的精度。一个三传感器目标跟踪系统的仿真例子说明其有效性。     

能量和带宽受限下的分布式一致性融合估计器

针对网络化多传感器分布式估计中传感器能量和通信网络带宽约束问题,提出一种基于降低发送频率和数据压缩降维的分布式一致性融合估计算法.为了满足通信网络带宽要求,各传感器节点直接选取局部估计信号的部分分量进行传输;与此同时,各节点随机间歇式发送数据包到其他节点来节省能量.在给定一致性权重下,建立以一致性估计器增益为决策变量,以所有传感器节点有限时域下状态融合估计误差协方差矩阵的迹的和为代价函数的优化问题,基于Lyapunov稳定性理论给出使得融合估计误差在无噪声时渐近稳定的一致性估计器增益存在的充分条件,并通过最小化代价函数的上界得到一组次优的一致性估计器增益值.最后,通过算例仿真验证算法的有效性.     

基于反馈结构的多传感器自适应航迹融合算法

在分布式多传感器信息融合系统中,自适应融合算法通过预先设定两个距离测度。然后将它们与逻辑判决树中的阈值进行比较来选择不同的融合算法,达到适应系统特性的不断变化,平衡精度与计算量之间矛盾的目的;反馈结构可用来提高局部节点的跟踪性能,进而提高全局航迹的融合精度．综合上述两种方法,提出一种新的基于反馈结构的多传感器自适应航迹融合算法,并在传感器异步的情况下讨论了其具体的计算过程．仿真结果表明该算法以较小的计算量达到了近似加权协方差算法（WCF）的融合精度．     

多模型多传感器信息融合Kalman平滑器

基于标量加权的线性最小方差最优信息融合算法,对多模型多传感器离散线性随机系统,给出了一种分布式标量加权信息融合固定滞后Kalman平滑器.它只需计算加权标量系数,可减小在融合中心的计算负担.当各子系统存在稳态滤波时,又给出了标量加权信息融合稳态平滑器,它计算量小,便于实时应用.并给出了两个子系统之间的平滑误差互协方差阵的计算公式.仿真例子验证了其有效性.     

基于带无偏输入的加权融合算法

给出了带输入估计的卡尔曼滤波算法,比较了带输入估计的卡尔曼滤波算法和标准卡尔曼滤波算法的差别,在此基础上提出了一种最优加权的数据融合方法,分析了最优权值的分配原则,给出了最优加权下的多传感器融合算法.该算法首先计算出各个传感器当前的滤波精度,依据各传感器当前时刻的滤波精度分配权值,同时测量方差的时变特性使得每次测量信息得到充分的利用,通过仿真比较了该方法与平均分配权值方法效果的差别,实验结果表明了该算法的有效性.     

一种权系数两级自调整的融合定位精度提高方法

在分布式传感器网络节点定位技术中,使用数据融合方法以提高探测系统的检测与定位精度正成为研究的热点。提出了一种应用于分布式传感器网络中的数据融合定位算法,通过对各个传感器节点的定位信息的加权求和来进行数据融合,用来提高探测系统目标定位的精度。该算法采用两级自适应调整得到最优加权因子,首先利用线性最小均方差（LMSE）算法得到权系数的初始值,然后利用训练节点和递归最小二乘（RLS）算法自适应地调整达到最优。对静态和运动目标的定位数据融合算法进行了仿真,仿真结果表明：相比单节点定位,提出的融合算法的定位精度有约1—2个数量级的提高。     

毫米波/红外复合制导信息融合技术研究

研究红外复合制导问题,对于复合制导中的集中式融合,与共同的噪声环境等因素使得各传感器的量测误差之间存相关性,且相关程度未知,对得到的融合估计不是最优的.红外导引头测量存在不完全可观性,一般的分布式融合也难以实现.针对以上问题,提出了一种基于协方差交叉法的信息融合算法,在估计相关性未知时利用信息空间上的均值和协方差凸组合,通过寻找行列式的最小方差阵,融合各传感器的局部估计,可以得到不低于单个传感器估计的融合估计精度.进行量测相关数据,并通过UKF滤波仿真验证,证明算法的有效性.     

Local node selection for target tracking based on underwater wireless sensor networks

Traditional sonar-array-based target tracking algorithms may be unsuitable for on-demand tracking missions, since they assume that the sonar arrays should be towed or mounted by a submarine or a ship. Alternatively, underwater wireless sensor networks can offer a promising solution approach. First, each underwater node is battery-powered, so saving energy expenditure is a critical issue. Instead of keeping all sensor nodes active, this paper provides a local node selection (LNS) scheme which increases energy efficiency by waking up only a small part of nodes at each time. Second, considering node's limited computing ability and the real-time requirement for the tracking algorithm, instead of employing the centralised fusion structure, we utilise the distributed Kalman filtering fusion with feedback in this paper. Finally, instead of assuming one sensor node can uniquely determine target's location, a more practical range-only measurement model is proposed. Then the LNS scheme and distributed fusion with feedback are extended to our range-only measurement model. The simulation results demonstrate the efficiency of our scheme.     

通信约束下的水声信号分布式检测融合算法研究

针对水声传感器网络中大延迟、低可靠通信约束下的水声信号分布式检测问题，提出了一种基于时间窗口的自适应融合算法．传感器节点依据声纳接收机的特性计算局部判决并发送给融合中心节点．融合中心节点在时间窗口内，基于已收到的局部判决在线自适应地调整融合规则，从而由最优融合算法得到最终判决．通过仿真，讨论了时间窗口的选择问题以及融合算法的性能．仿真结果表明，新算法具有很高的实用性，能够在动态变化的水声通信条件下保证整个系统高效运行．     

Distributed fusion filters from uncertain measured outputs in sensor networks with random packet losses

This paper addresses the distributed fusion filtering problem for discrete-time random signals from measured outputs perturbed by random parameter matrices and correlated additive noises. These measurements are obtained by a sensor network with a given topology, where random packet dropouts occur during the data transmission through the different network communication channels. The distributed fusion estimation is accomplished in two phases. Firstly, by an innovation approach and using the last observation that successfully arrived if a packet is lost, a preliminary distributed least-squares estimator is designed at each sensor node using its own measurements and those from its neighbors. Secondly, every sensor collects the preliminary filters that are successfully received from its neighbors and fuses this information with its own one to generate the least-squares linear matrix-weighted distributed fusion estimator. The accuracy of the proposed estimators, which is measured by the estimation error covariances, is examined by a numerical simulation example.     

Distributed EM Algorithm for Gaussian Mixtures in Sensor Networks

This paper presents a distributed expectation–maximization (EM) algorithm over sensor networks. In the E-step of this algorithm, each sensor node independently calculates local sufficient statistics by using local observations. A consensus filter is used to diffuse local sufficient statistics to neighbors and estimate global sufficient statistics in each node. By using this consensus filter, each node can gradually diffuse its local information over the entire network and asymptotically the estimate of global sufficient statistics is obtained. In the M-step of this algorithm, each sensor node uses the estimated global sufficient statistics to update model parameters of the Gaussian mixtures, which can maximize the log-likelihood in the same way as in the standard EM algorithm. Because the consensus filter only requires that each node communicate with its neighbors, the distributed EM algorithm is scalable and robust. It is also shown that the distributed EM algorithm is a stochastic approximation to the standard EM algorithm. Thus, it converges to a local maximum of the log-likelihood. Several simulations of sensor networks are given to verify the proposed algorithm.      

Distributed Kalman filtering via node selection in heterogeneous sensor networks

In this paper, we propose a strategy for distributed Kalman filtering over sensor networks, based on node selection, rather than on sensor fusion. The presented approach is particularly suitable when sensors with limited sensing capability are considered. In this case, strategies based on sensor fusion may exhibit poor results, as several unreliable measurements may be included in the fusion process. On the other hand, our approach implements a distributed strategy able to select only the node with the most accurate estimate and to propagate it through the whole network in finite time. The algorithm is based on the definition of a metric of the estimate accuracy, and on the application of an agreement protocol based on max-consensus. We prove the convergence, in finite time, of all the local estimates to the most accurate one at each discrete iteration, as well as the equivalence with a centralised Kalman filter with multiple measurements, evolving according to a state-dependent switching dynamics. An application of the algorithm to the problem of distributed target tracking over a network of heterogeneous range-bearing sensors is shown. Simulation results and a comparison with two distributed Kalman filtering strategies based on sensor fusion confirm the suitability of the approach.     

基于自适应加权融合的分布式滤波算法

针对存在丢包的传感器网络中每个传感器节点对目标估计确信度不同的问题,提出一种基于自适应加权融合的分布式滤波算法.考虑节点在网络中的影响力及其节点属性,将节点重要度与传感器网络节点观测数据间的支持度线性加权,获得每个传感器节点对目标的估计确信度,并将该确信度构成的融合权值引入节点状态估计值的一致性协议中,更新传感器节点对目标的状态估计值,提高分布式滤波算法的估计精度和传感器节点估计值的一致性.仿真结果验证了所提出方法的有效性.     

Self-organization of unattended wireless acoustic sensor networks for ground target tracking

A distributed, self-organization algorithm for ground target tracking using unattended acoustic sensor network is developed. Instead of using microphone arrays, each sensor node in the sensor network uses only a single microphone as its sensing device. This design can greatly reduce the size and cost of each sensor node and allow more flexible deployment of the sensor network. The self-organization algorithm presented in this paper can dynamically select proper sensor nodes to form the localization sensor groups that can work as a virtual microphone array to perform energy efficient target localization and tracking. To achieve this, we use a time-delay based bearing estimation plus triangulation for source localization in the sensor network. Major error sources of the localization method like time delay estimation, bearing calculation and triangulation are analyzed and sensor selection criteria are developed. Based on these criteria and neighborhood information of each sensor node, a distributed self-organization algorithm is developed. Simulation results show the effectiveness of the proposed algorithm.     

一种改进的无线多媒体传感器网络分布式图像压缩算法

针对无线多媒体传感器网络(WMSNs)中单个节点资源严重受限而难以直接处理和传输图像的问题,在分布式渐进图像压缩算法(DICA)的基础上,提出了一种改进的基于簇内分布式处理的图像压缩算法(ICDP),即在每个簇内通过能量优先选择原则选取辅助节点来协同完成JPEG2000图像压缩标准中的多级小波变换,从而实现图像的分布式处理和压缩。仿真结果表明,与DICA算法相比,ICDP算法能够在保证图像重建质量和压缩比相同的前提下,更好地平衡网络中各节点能耗,延长了网络的生命周期,更适合应用于资源受限、节点部署密集的WMSNs中。