基于高斯和与SCKF的非线性非高斯滤波算法

针对均方根容积卡尔曼滤波(SCKF)对非高斯情况滤波效果差的问题,在分析SCKF和高斯和滤波基础上,提出一种高斯和均方根容积卡尔曼滤波新算法。算法采用高斯和形式来逼近非高斯后验概率密度,将SCKF作为子滤波器,对每个高斯分量进行时间和量测更新,使其有效解决非线性非高斯滤波问题。仿真结果表明,高斯和均方根容积卡尔曼滤波估计精度高于粒子滤波和高斯和扩展卡尔曼滤波算法,与容积粒子滤波精度相当,但耗时约为容积粒子滤波的15%,是一种较好平衡跟踪精度和实时性的非线性非高斯滤波算法。     

非线性滤波方法及其在故障诊断中的应用

传统的滤波方法一般基于线性化和高斯假设,在一定程度上影响了滤波精度和非线性系统故障诊断的准确率。该文从"近似非线性"和"近似概率"的方法入手,分析3种常用的非线性滤波算法:扩展卡尔曼滤波器(EKF)、U-卡尔曼滤波器(UKF)以及粒子滤波器(PF)的原理、方法及特点并介绍其在非线性故障诊断中的应用价值。     

一种基于卡尔曼滤波的非线性模型在线模糊辨识算法

文章提出一种用于非线性模型在线辨识的模糊算法。该算法将非线性输入输出系统用时变线性系统模型来拟和，并把此非线性系统模型表示成模糊模型的形式，用在线调节模糊模型的方法来辨识时变线性模型的相关参数。本文将递推模糊聚类方法与卡尔曼滤波法用于在线调整模糊模型参数。仿真算例表明了此算法的有效性。     

一种用于光电目标跟踪的非线性滤波算法

基于传统转换测量卡尔曼滤波算法,提出了一种二阶去偏转化测量卡尔曼滤波算法。该算法对转换测量方程进行二阶泰勒展开,得到转换测量值误差的均值和方差表达式,并对转换测量方程进行去偏差补偿,再经转换测量卡尔曼滤波,从而显著减小传统滤波算法的线性化误差,提高远距离目标的跟踪精度。仿真结果表明:二阶去偏转换测量卡尔曼滤波(SCMKF)算法跟踪精度明显优于传统的扩展卡尔曼滤波(EKF)算法和转换测量卡尔曼滤波(CMKF)算法,收敛速度比EKF算法至少可提高1倍。     

多终端五阶强跟踪容积卡尔曼滤波实时定轨

为了提高多手持终端对卫星脉冲机动轨道的实时确定精度,并降低定轨精度对滤波初值的敏感性,提出一种五阶强跟踪球面单形-径向容积卡尔曼滤波(5-STSSRCKF)方法。采用n维正则单形变换群和矩匹配法推导五阶球面单形-径向容积准则,并将该准则嵌入强跟踪滤波(STF)框架。利用STF的等价表示计算次优渐消因子,实现了在滤波稳定时在线实时调整增益矩阵,克服了残差增大时增益矩阵仍保持极小值的缺点,提高了对系统突变状态的跟踪能力。进行了仿真实验,结果表明,当滤波初值误差增大时,已有方法的定轨精度降低为1 798.199m,而提出的5-STSSRCKF的定轨精度可以维持在8.688m;当卫星进行脉冲机动时,已有方法不具备对机动轨道状态的跟踪能力,而5-STSSRCKF的定轨精度仍然维持在8.976m。实验显示,5-STSSRCKF对滤波初值不敏感,并且对卫星脉冲机动轨道的实时确定精度更高。     

一种快速GPS/DR组合导航系统状态估计算法研究

无轨迹卡尔曼滤波(UKF)技术在非线性系统(GPS/DR车载组合导航系统)的状态估计中取得了比扩展卡尔曼滤波(EKF)更好的滤波精度和收敛速度.为了进一步减少采样点数目,提高UKF滤波实时性,一组n+2个采样点被构造用于逼近系统状态分布.蒙特卡洛仿真表明RUKF和UKF在滤波精度和收敛速度上是一致的,RUKF的计算效率好于UKF.     

一种小型无人机无源目标定位方法及精度分析

为了提高小型无人机无源目标定位的精度,设计了一种新的目标定位算法。首先确定目标定位过程中的坐标转换关系并推导出视轴角的计算模型;然后,利用光电侦察平台锁定跟踪目标的特性,提出了对同一目标点多次测量的目标定位框架,建立了系统状态方程和测量方程,考虑到测量方程的非线性,将无迹卡尔曼滤波应用于目标位置估计;最后,针对加性高斯白噪声的非线性目标定位系统,推导出理论上的定位误差的克拉美-罗下限。仿真结果表明,该算法具有较高的目标定位精度,滤波器估计误差均方差已逼近非线性系统的克拉美-罗下限。现场试验结果表明,在离地面约1000 m的空中,无人机对地面目标定位精度可达8.1 m。该算法易于部署,可操作性强,具有较大的实用价值。     

室内全自主智能吸尘器的一种定位方法

针对室内全自主智能吸尘器,建立了二维运动学模型和以超声传感为主的多传感器融合,并采用融合区间分析算法的扩展卡尔曼滤波处理方法,作了对比仿真实验。结果表明,这种定位方法效果优良,可以成功运用在智能吸尘器上。     

一种退役锂电池健康状态估计方法

针对退役锂电池健康状态估计效率较低的现状,提出一种快速、有效的估计方法。首先采用3阶RC等效电路模型描述电池特性得出状态方程,确保电池模型精确性,同时引入电池荷电状态SOC(State of charge)和欧姆内阻(R₀)作为状态方程参数。其次利用区域概念,计算出特定的区域容量与区域电压,减少电池参数估计所需要的数据、时间。然后通过扩展卡尔曼滤波(Extended kalman filtering)算法估计电池参数SOC和R₀,进而对电池健康状态(State of health, SOH)进行估计。最后,利用电池测试设备(Arbin-BT2000)对18650电池进行充放电实验,验证该方法的可行性。实验结果证明SOH估计所需参数明显减少,使得电池数据测量所需时间明显缩短,并且估计误差不超过4%,误差较小,说明所提出方法能快速、有效地估算出电池SOH。     

一种新型数字伺服油缸的非线性仿真分析

建立一种用伺服螺旋机构原理设计的小行程数字步进伺服油缸的动态数学模型。基于非线性仿真模块,在MAT-LAB/Simulink环境中,根据活塞下腔长度、工作负荷、螺旋槽断面尺寸和活塞半径四个主要结构及工作参数的变化,获得动态响应曲线,并仿真分析该数字伺服油缸的频率特性。结果表明,该数字伺服油缸在满负载和小行程工况下具有较好的动态特性,可广泛应用于数控调整系统领域。     

基于卡尔曼滤波的动态传感数据流估计方法

在无线传感器网络应用系统中,众多传感器节点以一定的时间间隔不断采集被监测对象的参数,对于数据中心而言,形成了无线传感数据流.同时数据流中的数据模型可能随时间变化,因此形成了动态的传感数据流.针对目前无线传感数据流估计方法中存在的估计精度较低、模型更新不及时和计算复杂度高等问题,提出了基于卡尔曼滤波的动态无线传感数据流估计方法.采用卡尔曼滤波实现估计模型的动态调整,为了降低数据流估计的计算复杂度,采用基于相关分析的多元线性回归估计方法,将卡尔曼滤波和多元线性回归模型有机结合,实现对动态无线传感数据流的准确估计.采用实际传感数据的估计实验结果表明,提出的基于卡尔曼滤波的动态传感数据流估计方法可有效实现动态传感数据流的估计.     

Novel method for evaluating surface roughness by grey dynamic filtering

The evaluation of surface roughness is of great importance for manufacturing industries as the roughness of a surface has a considerable influence on its quality and the function of products. For surface roughness evaluation, to find an appropriate reference line is of the utmost importance. A smooth grey reference line obtained by grey dynamic filtering is proposed to evaluate surface roughness. The primary sampling data of the measured surface need not obey the typical distributions and the surface profile with less data can also be evaluated without losing primary data. Through sample analysis, the grey reference line is well consistent with ISO Gaussian reference line and their evaluation results for surface roughness are in agreement. The grey reference line can be used as one of complements for Gaussian reference line.     

Combined extended FIR/Kalman filtering for indoor robot localization via triangulation

A combined unbiased finite impulse response (UFIR) and Kalman filtering algorithm is proposed for mobile robot localization via triangulation utilizing noisy measurements. We consider a mobile robot travelling on an indoor floorspace with three nodes in a view. Under the not well-known initial robot state and noise statistics, the extended Kalman filter (EKF) may produce unacceptable estimates. The iterative extended UFIR (EFIR) filter ignores the noise statistics, but requires N initial points of linear measurements which are unavailable. The combined EFIR/Kalman algorithm utilizes N first EKF estimates with approximately set initial conditions and noise statistics as linear measurements for EFIR filter. It is shown that the combined algorithm is more accurate than EKF in robot localization under the real operation conditions. Simulations are provided for piecewise and circular robot trajectories.     

Railway track geometry determination using adaptive Kalman filtering model

The Kalman filter has been widely used to solve different filtering problems especially in tracking and estimation applications. Besides its simplicity, robustness and optimality, the application of Kalman filter to nonlinear systems can be complicated. The most common method is to use extended Kalman filter which linearizes the nonlinear model so that the standard Kalman filter can be applied. In this paper, a new adaptive Kalman filtering algorithm is designed and applied to a railway track geometry surveying system which has been designed in the scope of a research project at Yildiz Technical University/Turkey. Track gauge, super-elevation, gradient and track axis coordinates which are the railway geometrical parameters can be instantly determined while making measurements by using adaptive Kalman filtering algorithm integrated surveying system.     

Comparison of Linearized Kalman Filter and Extended Kalman Filter for Satellite Motion States Estimation

The performance of the conventional Kalman filter depends on process and measurement noise statistics given by the system model and measurements.The conventional Kalman filter is usually used for a linear system,but it should not be used for estimating the state of a nonlinear system such as a satellite motion because it is difficult to obtain the desired estimation results.The linearized Kalman filtering approach and the extended Kalman filtering approach have been proposed for a general nonlinear system.The equations of satellite motion are described.The satellite motion states are estimated,and the relevant estimation errors are calculated through the estimation algorithms of the both above mentioned approaches implemented in Matlab are estimated.The performances of the extended Kalman filter and the linearized Kalman filter are compared.The simulation results show that the extended Kalman filter is much better than the linearized Kalman filter at the aspect of estimation effect.     

故障诊断中非线性滤波算法的研究

对扩展卡尔曼滤波(EKF)和粒子滤波(PF)的原理进行了介绍。针对故障诊断中的非线性非高斯问题,通过仿真实验比较了EKF和PF的效果,结果证明在非线性条件下,PF的算法优于EKF算法。     

一种基于Kalman滤波算法的组合式温度传感器的研究

该文研究了一种基于Kalman滤波算法的组合式温度传感器。根据铂电阻和半导体热敏电阻在温度测量中的不同特性,设计了一种组合式温度传感器,并利用Kalman滤波算法进行综合数据处理。提出了基于Kalman滤波算法的组合式温度传感器模型,给出了静态测量和动态测量两种情况下的kalman滤波算法步骤,分析了Kalman滤波算法的参数设置。实验结果表明,该组合式温度传感器可有效提高测量结果的准确性和灵敏度。     

卡尔曼滤波预测在客流量统计中的应用

运动物体的跟踪是现代科学技术里的重要一环,在众多领域都有着广泛的应用,比如说雷达检测跟踪系统。运动目标的跟踪分为预测和匹配两方面,其中预测尤为重要。文中运用向量卡尔曼滤波的方法对在5W红外LED光源照射下的人体的运动进行实时跟踪预测,为后续匹配工作提供高准确率的保障。在实验室模拟环境下进行实验,实验结果表明预测的误差能够控制在10个像素点以内。在对公交车客流视频进行处理后,有非常高的准确率,实时性也很好。     

Adaptive Vold–Kalman filtering order tracking

This study proposes and implements an adaptive Vold–Kalman filtering order tracking (VKF_OT) approach to overcome the drawbacks of the original VKF_OT scheme for condition monitoring and diagnosis of rotary machinery. The paper comprises theoretical derivation and numerical implementation. Comparisons of the adaptive VKF_OT scheme to the original are accomplished through processing two synthetic signals composed of close orders and crossing orders, respectively. Parameters such as the weighting factor and the correlation matrix of process noise, which influence tracking performance, are investigated. The adaptive scheme coping with end effects of computation can simultaneously extract multiple order/spectral components, and effectively decouple close and/or crossing orders associated with multi-axial reference rotating speeds. Furthermore, the adaptive OT scheme is realized through Kalman filtering based upon adapted one-step prediction scheme, where a parameter, the weighting factor, is newly introduced into the computation. Thus the technique can be computed on-line and implemented as a real-time processing application.