Stability and performance analysis of the compressed Kalman filter algorithm for sparse stochastic systems

This paper considers the problem of estimating unknown sparse time-varying signals for stochastic dynamic systems.To deal with the challenges of extensive sparsity,we resort to the compressed sensing method and propose a compressed Kalman filter(KF)algorithm.Our algorithm first compresses the original high-dimensional sparse regression vector via the sensing matrix and then obtains a KF estimate in the compressed low-dimensional space.Subsequently,the original high-dimensional sparse signals can be well recovered by a reconstruction technique.To ensure stability and establish upper bounds on the estimation errors,we introduce a compressed excitation condition without imposing independence or stationarity on the system signal,and therefore suitable for feedback systems.We further present the performance of the compressed KF algorithm.Specifically,we show that the mean square compressed tracking error matrix can be approximately calculated by a linear deterministic difference matrix equation,which can be readily evaluated,analyzed,and optimized.Finally,a numerical example demonstrates that our algorithm outperforms the standard uncompressed KF algorithm and other compressed algorithms for estimating high-dimensional sparse signals.

Research on Kalman Filtering Algorithmfor Deformation Information Series ofSimilar Single-Difference Model

Using similar single-difference methodology(SSDM) to solve the deformation values of the monitoring points, there is unstability of the deformation information series, at sometimes.In order to overcome this shortcoming, Kalman filtering algorithm for this series is established,and its correctness and validity are verified with the test data obtained on the movable platform in plane. The results show that Kalman filtering can improve the correctness, reliability and stability of the deformation information series.

Multi-sensor Hybrid Fusion Algorithm Based on Adaptive Square-root Cubature Kalman Filter

In the normal operation condition, a conventional square-root cubature Kalman filter （SRCKF） gives sufficiently good estimation results. However, if the measurements are not reliable, the SRCKF may give inaccurate results and diverges by time. This study introduces an adaptive SRCKF algorithm with the filter gain correction for the case of measurement malfunctions. By proposing a switching criterion, an optimal filter is selected from the adaptive and conventional SRCKF according to the measurement quality. A subsystem soft fault detection algorithm is built with the filter residual. Utilizing a clear subsystem fault coefficient, the faulty subsystem is isolated as a result of the system reconstruction. In order to improve the performance of the multi-sensor system, a hybrid fusion algorithm is presented based on the adaptive SRCKF. The state and error covariance matrix are also predicted by the priori fusion estimates, and are updated by the predicted and estimated information of subsystems. The proposed algorithms were applied to the vessel dynamic positioning system simulation. They were compared with normal SRCKF and local estimation weighted fusion algorithm. The simulation results show that the presented adaptive SRCKF improves the robustness of subsystem filtering, and the hybrid fusion algorithm has the better performance. The simulation verifies the effectiveness of the proposed algorithms.

TEC and Instrumental Bias Estimation of GAGAN Station Using Kalman Filter and SCORE Algorithm

The standalone Global Positioning System (GPS) does not meet the higher accuracy requirements needed for approach and landing phase of an aircraft. To meet the Category-I Precision Approach (CAT-I PA) requirements of civil aviation, satellite based augmentation system (SBAS) has been planned by various countries including USA, Europe, Japan and India. The Indian SBAS is named as GPS Aided Geo Augmented Navigation (GAGAN). The GAGAN network consists of several dual frequency GPS receivers located at various airports around the Indian subcontinent. The ionospheric delay, which is a function of the total electron content (TEC), is one of the main sources of error affecting GPS/SBAS accuracy. A dual frequency GPS receiver can be used to estimate the TEC. However, line-of-sight TEC derived from dual frequency GPS data is corrupted by the instrumental biases of the GPS receiver and satellites. The estimation of receiver instrumental bias is particularly important for obtaining accurate estimates of ionospheric delay. In this paper, two prominent techniques based on Kalman filter and Self-Calibration Of pseudo Range Error (SCORE) algorithm are used for estimation of instrumental biases. The estimated instrumental bias and TEC results for the GPS Aided Geo Augmented Navigation (GAGAN) station at Hyderabad (78.47°E, 17.45°N), India are presented.

Assimilation of Remote Sensing and Crop Model for LAI Estimation Based on Ensemble Kalman Filter

Data assimilation in agricultural remote sensing research is of great significance to integrate with remote sensing observations and model simulations for parameters estimation. The present investigation not only designed and realized the Ensemble Kalman Filtering algorithm （EnKF） assimilation by combing the crop growth model （CERES-Wheat） with remote sensing data, but also optimized and updated the key parameters （LAI） of winter wheat by using remote sensing data. Results showed that the assimilation LAI and the observation ones agreed with each other, and the R2 reached 0.8315. So assimilation remote sensing and crop model could provide reference data for the agricultural production.

基于CamShift与Kalman相结合的目标跟踪算法研究

目标跟踪是机器视觉领域的一项重要技术。传统的CamShift目标跟踪算法具有时间复杂度低、运算速度快的优点,在简单背景下具有良好的跟踪效果。但当跟踪目标处于部分遮挡的复杂情况下时,容易出现目标丢失的情况,从而影响后续的跟踪。Kalman滤波算法在目标跟踪任务中,能够有效地预测跟踪目标下一时刻可能出现的位置,且算法简单方便。现将CamShift算法与Kalman滤波算法相结合来优化传统的CamShift算法,实验结果表明:优化后的算法不但保留了传统CamShift算法的优点,而且在一定程度上可以预测跟踪目标的行动轨迹,更好地实现目标跟踪任务,解决了传统CamShift算法在目标被部分遮挡的情况下容易出现的目标丢失的情况;同时,运算结果准确,基本没有预测误差。

Real-time localization estimator of mobile node in wireless sensor networks based on extended Kalman filter

Localization of the sensor nodes is a key supporting technology in wireless sensor networks （WSNs）. In this paper, a real-time localization estimator of mobile node in WSNs based on extended Kalman filter （KF） is proposed. Mobile node movement model is analyzed and online sequential iterative method is used to compute location result. The detailed steps of mobile sensor node self-localization adopting extended Kalman filter （EKF） is designed. The simulation results show that the accuracy of the localization estimator scheme designed is better than those of maximum likelihood estimation （MLE） and traditional KF algorithm.

基于卡尔曼(Kalman)滤波算法的体育训练视频中的运动目标检测

体育训练视频中的运动目标检测往往存在目标快速移动、目标遮挡和场景变化等问题,导致运动目标检测的难度增大。为了解决这些问题,提出了基于卡尔曼(Kalman)滤波算法的体育训练视频中运动目标检测的方法。为了提升体育训练视频图像的质量,首先对体育训练视频的图像进行一系列预处理(包括灰度变换、轮廓对比增强和噪声抑制等);然后借助混合高斯模型(Gaussian mixture module,GMM)有效提取体育训练视频的前景信息;为了精准捕捉体育训练视频中的运动目标,运用三帧差分法设定Kalman滤波器的初始状态,利用高效检测算法准确获取每一帧图像中运动目标的观测位置;随后将初始状态和观测位置的数据输入Kalman滤波器;最后在Kalman滤波器中,结合上一帧的预估值和当前帧的监测值,对体育训练视频中运动目标的当前状态进行精确估算和优化,并对下一帧的状态进行预测,从而实现了对运动目标的持续跟踪与精准预测。实验结果表明,与基于特征融合的全卷积孪生网络(siamese full convolution,Siamfc)目标追踪算法和基于连续自适应均值漂移(continuously adapting mean shift,Camshift)的均值漂移(mean shift,Meanshift)改进算法相比,采用基于Kalman滤波算法的运动目标检测方法对运动目标的重叠精度(overlap precision,OP)和中心位置误差(center location error,CLE)进行检测,不仅能够有效检测到体育训练视频中的运动目标,还可表现出较高的准确性和实时性。

模糊自适应Kalman滤波算法在SINS/DR组合导航的应用

针对里程仪测量误差导致组合导航精度降低的问题,提出基于系统工作状态和滤波器新息状态相结合的模糊自适应算法。根据新息的变化确定模糊规则,修正里程仪输出增益,使新息始终保持在零均值附近,利用修正后的新息修正观测噪声方差,降低导航定位的偏差。仿真实验结果证明,该算法能够很好地提高组合导航定位的精度。

Kalman算法在纯电动汽车SOC估算中的应用误差分析

针对纯电动汽车电池组的工作状态和输出特性,分析了模型参数的变化对Kalman算法估算精度的影响。指出了纯电动汽车应用Kalman滤波算法估算SOC应考虑的因素,并结合电池模型参数的变化提出了Kal-man方程修正方案。最后通过电池的城市工况模拟试验,验证了分析的正确和可行性。

残膜回收机有效作业面积监测系统设计研究

为便于新疆棉田残膜回收的精准作业收费和补贴,设计了一种残膜回收机有效作业面积监测系统。通过仿真试验对比锁相放大器降噪前后残膜回收机作业时经纬度数据,得出锁相滤波对于噪声环境有较好的抗干扰性能;采用高斯投影正算方法将降噪后经纬度数据转化为平面坐标,得出1.5°带的高斯投影的相对误差相较于3°带降低了0.01%。同时,利用卡尔曼滤波算法和构成角方法筛选有效作业轨迹,使用缓冲区法计算重复作业面积,最终利用轨迹算法精准获取残膜回收机有效作业面积,得出所设计的残膜回收机有效作业面积监测系统对于大块农田测量时相对误差为2.6%,对于小块农田测量时相对误差为3.3%。研究结果表明:所设计的残膜回收机有效作业面积监测系统可以精确计算出残膜回收机的有效作业面积,为残膜回收机作业质量评估提供了一定的理论基础。

一种快速Kalman滤波算法实现及效果评估

该文介绍了一种新的快速Kalman滤波算法实现方法。对于某些不能够采取离线计算的滤波过程来说,它可以在保证一定精度的同时极大地提高计算速度和减少计算占用资源。文中以仿真实验的跟踪数据做出了对比仿真。

时变系统的Laguerre模型辨识及设计变量(2)——Kalman滤波法

文章考虑动态线性系统的时变参数是平稳的AR(1)变量,系统为时变的Laguerre模型时的传递函数估计的均方误差(MSE)。在缓慢时变和高阶模型下,利用Kalman滤波算法,得到MSE的近似表达式。最后得到了Kalman滤波算法的设计变量的最优解。

HSV颜色空间特征与Kalman滤波融合的目标跟踪

为了克服噪声、遮挡、背景的改变等对目标识别带来的困难,出现了很多的跟踪算法。提出了一种基于HSV色彩空间的目标跟踪融合算法,即在较短时间内,将目标的运动看作一时不变系统,引入卡尔曼滤波进行参数辨识,使得跟踪系统具有后续状态预测的能力。算法包括均值漂移算法跟踪下利用卡尔曼滤波对后续状态预测和基于卡尔曼滤波状态估计的Bhattacharyya系数分析两个子过程,整个跟踪过程分两个子过程交替执行。对不同的视频序列测试的结果表明,算法能够对目标进行持续、稳健的跟踪。验证了新方法的有效性和准确性。

Kalman滤波融合优化Mean Shift的目标跟踪算法

目标跟踪中,目标的背景变化、形状改变、遮挡,往往会导致跟踪失败,而跟踪的实时性和准确性是必须考虑的问题。本文首先对Mean Shift算法进行了介绍,接着对Mean Shift算法进行了优化:修正Mean Shift算法迭代权值,修正后主要信息贡献更加突出,次要信息受到抑制,避免了开方的繁琐运算,降低了运算量。提出了目标模板更新算法,解决了背景变化和目标形状改变时跟踪失败的问题。然后在水平位置和竖直位置建立Kalman滤波器,同时将优化Mean Shift算法与Kalman滤波融合,解决了目标完全遮挡后无法继续跟踪的问题。仿真实验表明,本文提出的目标跟踪算法在目标遮挡,目标形状改变,目标跟踪失败的情况下具有更高的跟踪精度,更高的实时性和鲁棒性。

结合SURF与Kalman滤波的CAMShift跟踪算法

针对传统的CAMShift目标跟踪算法,在出现颜色干扰,遮挡等复杂背景中容易跟丢的问题,提出了一种结合SURF特征匹配与Kalman滤波的CAMShift跟踪算法。该算法利用CAMShift算法跟踪得到的候选目标与模板目标的色度和梯度方向的综合直方图比较计算得到的Bhattacharyya系数作为判定依据,当系数大于给定阈值时,采用SURF算法对搜索窗口和上一帧跟踪结果进行特征匹配,重新计算目标的大小和位置。同时为了避免目标快速运动时跟踪失败和减少SURF匹配的计算量,利用Kalman滤波对运动目标窗口进行预测更新以确定下一帧搜索窗口的中心位置。实验表明,该算法在图像背景复杂,出现颜色干扰以及部分遮挡时能够稳定跟踪,其跟踪速度与结合SURF的CAMShift算法相比有显著提高。

成长期上市公司财务危机预测——基于Kalman滤波与Logistic回归的实证研究

本文将Kalman滤波智能算法与Logistic回归传统模型相结合,对成长期上市公司财务危机进行预测。结果表明:公司治理因素对上市公司是否发生财务危机具有显著影响;比较而言,Kalman滤波算法算得的专一性优于Logistic回归模型;临近被ST的T-1期模型Ⅰ、Ⅱ以及T-3期模型Ⅱ计算所得敏感性高于Kalman滤波算法敏感性结果,但T-3期模型Ⅰ及T-5期模型Ⅰ、Ⅱ敏感性皆低于Kalman滤波算法得到的敏感性。文章最后提出了成长期上市公司避免陷入财务困境的政策建议。

基于改进Mean-Shift与自适应Kalman滤波的视频目标跟踪

提出一种改进的Mean-Shift和自适应Kalman滤波器相结合的视频运动目标跟踪算法。对选定的跟踪目标,采用三帧差和区域增长法分割目标并得到主颜色信息。在跟踪过程中,利用自适应的Kalman滤波器估计每一帧的起始迭代位置,再利用改进的Mean-Shift算法得到跟踪位置并作为测量值反馈给自适应Kalman滤波器,并引入遮挡率因子以自适应地调节Kalman估计参数。实验结果表明,该算法能对视频中的运动目标实现检测和连续跟踪,对遮挡也有较好的鲁棒性。

基于两次Kalman滤波的观测噪声自适应调整算法

在实际的目标跟踪过程中,由于目标远近等各种客观因素的影响,观测噪声是随时变化的。但是在标准卡尔曼滤波中,如果将观测噪声协方差设为恒定值,必然造成跟踪结果不理想。针对这种情况,通过在任意时刻施行两次卡尔曼滤波的结果来自适应地调整观测噪声协方差,使卡尔曼滤波算法中的观测噪声协方差与实际值更加接近,从而提高对目标的跟踪精度。最后Monte Carlo仿真实验证明了本算法的有效性。

基于强跟踪Kalman滤波的鲁棒人脸跟踪算法

进行人脸跟踪时,人脸出现快速移动、姿态大幅偏转及被遮挡的情况会导致人脸跟踪失败。针对以上问题,结合Haar-like特征与Adaboost算法对正面人脸进行跟踪,当人脸姿态出现大幅度偏转或被遮挡时,使用人脸肤色的直方图特征与Camshift算法对人脸进行跟踪。为提高人脸跟踪算法的运行效率,使用强跟踪Kalman滤波算法对人脸在下一帧图像中可能出现的位置进行预测,缩小人脸检测的搜索区域,提高人脸跟踪算法的实时性。实验验证了该算法在视频人脸跟踪方面具有较好的鲁棒性和实时性。