Multisensor Data Fusion for High Quality Data Analysis and Processing in Measurement and Instrumentation

Multisensor data fusion （MDF） is an emerging technology to fuse data from multiple sensors in order to make a more accurate estimation of the environment through measurement and detection. Applications of MDF cross a wide spectrum in military and civilian areas. With the rapid evolution of computers and the proliferation of micro-mechanical/electrical systems sensors, the utilization of MDF is being popularized in research and applications. This paper focuses on application of MDF for high quality data analysis and processing in measurement and instrumentation. A practical, general data fusion scheme was established on the basis of feature extraction and merge of data from multiple sensors. This scheme integrates artificial neural networks for high performance pattern recognition. A number of successful applications in areas of NDI （Non-Destructive Inspection） corrosion detection, food quality and safety characterization, and precision agriculture are described and discussed in order to motivate new applications in these or other areas. This paper gives an overall picture of using the MDF method to increase the accuracy of data analysis and processing in measurement and instrumentation in different areas of applications.

Research on Kalman-filter based multisensor data fusion

Multisensor data fusion has played a significant role in diverse areas ranging from local robot guidance to global military theatre defense etc. Various multisensor data fusion methods have been extensively investigated by researchers, of which Klaman filtering is one of the most important. Kalman filtering is the best-known recursive least mean-square algorithm to optimally estimate the unknown states of a dynamic system, which has found widespread application in many areas. The scope of the work is restricted to investigate the various data fusion and track fusion techniques based on the Kalman Filter methods, then a new method of state fusion is proposed. Finally the simulation results demonstrate the effectiveness of the introduced method.

Multisensor image fusion algorithm using nonseparable wavelet frame transform

A muitisensor image fusion algorithm is described using 2-dimensional nonseparable wavelet frame （NWF） transform. The source muitisensor images are first decomposed by the NWF transform. Then, the NWF transform coefficients of the source images are combined into the composite NWF transform coefficients. Inverse NWF transform is performed on the composite NWF transform coefficients in order to obtain the intermediate fused image. Finally, intensity adjustment is applied to the intermediate fused image in order to maintain the dynamic intensity range. Experiment resuits using real data show that the proposed algorithm works well in muitisensor image fusion.

THE RESEARCH OF GRADATION FUSION ALGORITHM BASED ON MULTISENSOR ASYNCHRONOUS SAMPLING SYSTEM

This letter explores the distributed multisensor dynamic system, which has uniform sampling velocity and asynchronous sampling data for different sensors, and puts forward a new gradation fusion algorithm of multisensor dynamic system. As the total forecasted increment value between the two adjacent moments is the forecasted estimate value of the corresponding state increment in the fusion center, the new algorithm models the state and the forecasted estimate value of every moment. Kalman filter and all measurements arriving sequentially in the fusion period are employed to update the evaluation of target state step by step, on the condition that the system has obtained the target state evaluation that is based on the overall information in the previous fusion period. Accordingly, in the present period, the fusion evaluation of the target state at each sampling point on the basis of the overall information can be obtained. This letter elaborates the form of this new algorithm. Computer simulation demonstrates that this new algorithm owns greater precision in estimating target state than the present asynchronous fusion algorithm calibrated in time does.

A NOVEL REGION FEATURE USED IN MULTISENSOR IMAGE FUSION

A new region feature which emphasized the salience of target region and its neighbors is proposed. In region segmentation-based multisensor image fusion scheme, the presented feature can be extracted from each segmented region to determine the fusion weight. Experimental results demonstrate that the proposed feature has extensive application scope and it provides much more information for each region. It can not only be used in image fusion but also be used in other image processing applications.

Multiscale Multisensor Data Fusion and Application in High Precision Marking and Cutting Robot System

The multisensor online measure system for high precision marking and cutting robot system is designed and the data fusion method is introduced, which combines augment state multiscale process with extend Kalman filter. The technology measuring the three-dimensional deforming information of profiled bars is applied. The experimental result shows that applying the multisensor data fusion technology can enhance the measure precision and the reliability of measure system.

THEORETICAL ANALYSIS OF IMPROVEMENT OF TRACK LOSS IN CLUTTER WITH MULTISENSOR DATA FUSION

The paper analyses the improvement of track loss in clutter with multisensor data fusion.By a determination of the transition probability density function for the fusion prediction error, one can study the mechanism of track loss analytically. With nearest-neighbor association algorithm. The paper we studies the fused tracking performance parameters, such as mean time to lose fused track and the cumulative probability of lost fused track versus the normalized clutter density, for track continuation and track initiation, respectively. A comparison of the results obtained with the case of a single sensor is presented. These results show that the fused tracks of multisensor reduce the possibility of track loss and improve the tracking performance. The analysis is of great importance for further understanding the action of data fusion.

Fuzzy Stochastic Approach for Multisensor Fusion

The problem of multisensor fuzzy stochastic fusion is probed in the paper. The concept of fuzzy stochastic fusion entropy is defined, the character of fusion entropy is discussed and the entropy rule of optimal decision in multisensor system is deduced first. The criterion of multisensor fuzzy stochastic data fusion is presented, and the adaptive algorithms of multisensor fuzzy random data fusuion under the criterion is set up second. The effectiveness of the decision fusion and data fusion method has been demonstrated through the computer simulation last.

A new multisensor fusion SLAM approach for mobile robots

This paper presents a novel method, which enhances the use of external mechanisms by considering a multisensor system, composed of sonars and a CCD camera. Monocular vision provides redundant information about the location of the geometric entities detected by the sonar sensors. To reduce ambiguity significantly, an improved and more detailed sonar model is utilized. Moreover, Hough transform is used to extract features from raw sonar data and vision image. Information is fused at the level of features. This technique significantly improves the reliability and precision of the environment observations used for the simultaneous localization and map building problem for mobile robots. Experimental results validate the favorable performance of this approach.

优化的ID3算法在多传感器安防系统中的应用

针对实物保护系统(Physical Protection System,PPSY)中单一传感器报警准确率较低的问题,提出了一种基于改进ID3的CAC-ID3(Confidence And Correlation-ID3)算法在多传感器实物保护系统中数据融合的新方法。与传统的单一传感器数据信息处理相比,多传感器数据融合能够更加准确、全面的得到被测对象的数据信息,有效地利用多传感器资源。CAC-ID3算法首先在ID3的基础上引入属性置信度重新计算期望熵,解决属性和价值不对等的问题,克服多传感器数据分类时多值偏向的缺点,其值由经验和相关领域知识决定。然后通过引入属性间的相关度来调整信息增益值,提高分类精度。实验结果表明:基于CAC-ID3的决策树算法的多传感器PPSY能有效提高报警准确率和可靠性,防止敌对分子入侵,提高传感器对PPSY的检测的效能,且该算法的分类精度高于ID3算法。

基于Bi-TCN-LSTM的滚动轴承剩余使用寿命预测方法

由于时间卷积网络(temporal convolutional networks, TCN)感知场不足,轴承的关键退化信息常常被忽略,导致轴承剩余使用寿命(remaining useful life, RUL)预测结果不佳;而长短期记忆网络(long short-term memory, LSTM)随着数据量及序列长度的增加,长期依赖问题仍可能得不到很好解决。因此,课题组提出了一种基于双向时间卷积网络和长短期记忆(Bi-TCN-LSTM)的滚动轴承寿命预测方法。首先对多传感器数据进行归一化并做融合处理,然后采用Bi-TCN-LSTM进行数据特征提取与深度学习,其中对TCN模块引入卷积注意力机制(convolutional attention module, CAM),将LSTM的3个门简化为1个门,有效加快了预测模型学习的速度并提高了预测模型的精确度;采用IEEE PHM 2012轴承数据集作为实验数据集,进行了RUL预测实验。结果表明:与其他先进的预测模型相比,Bi-TCN-LSTM方法预测结果的误差相对较低,预测性能较好。

基于MSF的煤矿井下环境信息危险评价系统的研究

文章结合煤矿井下环境信息的特征,介绍了一种基于多传感器信息融合(MSF)的煤矿井下环境信息危险评价系统。该系统建立了用于煤矿井下环境信息危险预测的3层误差反向传播神经网络模型,并采用神经网络信息融合算法对样本数据进行了分析和处理。仿真结果表明,该系统能够比较准确地评价煤矿井下环境危险的程度,且具有较好的鲁棒性和泛化能力。

基于MSF技术的汽轮发电机状态估计

该文提出一个有效的基于径向基函数神经网络的模型和状态数据融合的汽轮发电机智能估计方法。文中阐述了其网络结构、学习算法、特征提取及综合决策方法。该模型同时利用了故障样本及专家经验知识,并通过不断学习新的样本获取新的知识,模型将越来越完善。仿真结果表明,该网络模型和信息融合方法是可行和有效的。

基于红外与激光雷达融合的鸟瞰图空间三维目标检测算法

结合MEMS激光雷达和红外相机的优势,设计了一种简单轻量、易于扩展、易于部署的可分离融合感知系统实现三维目标检测任务,将激光雷达和红外相机分别设置成独立的分支,两者不仅能独立工作也能融合工作,提升了模型的部署能力。模型使用鸟瞰图空间作为两种不同模态的统一表示,相机分支和雷达分支分别将二维空间和三维空间统一到鸟瞰图空间下,融合分支使用门控注意力融合机制将来自不同分支的特征进行融合。通过实际场景测试验证了算法的有效性。

多传感器融合的无人车自主定位实验研究

为了满足机器人专业课程实验研究性教学需求,设计一个多传感器融合的无人车自主定位实验。选择智能机器人开放平台作为载体,在硬件平台上研究退化环境实验,实现3D激光惯性融合的定位与建图。在实验环节中,首先,通过惯性测量单元(IMU)获得位姿信息,通过激光雷达获得点云数据,利用扩展卡尔曼滤波处理位姿信息,并利用体素滤波处理点云数据,从而完成数据预处理;然后通过坐标转换实现激光惯性组合定位;最后在硬件平台上研究退化环境实验,实现3D激光惯性融合的定位与建图。实验结果表明,IMU测量数据有较高的准确性,而激光点云则会约束IMU的测量偏差。这种组合方式能够有效地提高同步定位与建图(SLAM)系统在复杂现实环境中的测量精度和鲁棒性,适用于无人车的自主定位。应用表明,该实验系统使学生获得了综合训练,提升了学生的综合实践创新能力和科研创新能力,并取得了良好的教学效果。

基于多传感信息融合的电力变压器绕组故障诊断方法

传统电力变压器绕组故障诊断方法直接对变压器绕组故障进行诊断,未对振动信号进行特征提取,造成诊断精准度低。为此,提出基于多传感信息融合的电力变压器绕组故障诊断方法。在故障诊断之前采集绕组振动信号,在采集绕组振动信号的基础上,重点对变压器绕组振动特征进行提取,通过提取的特征利用多传感信息融合算法对变压器绕组变形故障进行故障诊断。设计对比实验,实验结果证明该方法在不同数据预处理下的故障诊断精准度高于传统方法,有一定的研究价值。

基于多传感器信息融合的光伏微电网设备故障预警研究

当前对于光伏微电网设备故障预警多采用相似性建模方法,但这种方法缺少对设备运行数据的流处理,导致故障预警精度较低。因此,提出基于多传感器信息融合的光伏微电网设备故障预警研究。首先,采集设备的状态参数,并利用多变量非线性函数对设备运行的历史数据与实时数据进行流处理与故障引擎分析,由此建立设备状态模型。其次,结合多传感器信息融合技术融合设备状态数据,并将其转化为后验概率输出,求取故障预警指标。最后,构建设备故障预警模型,计算故障预警区间阈值,实现设备故障预警。结果表明:所提方法得到的预警值与真实值之间的拟合优度较高,具有较高的预警精度。

面向硫磺生产线的粉尘适应性AGV系统设计

在硫磺生产线的特殊环境下,尤其是硫磺粉尘对自动化引导车辆(AGV)系统功能的影响,为自动化运输系统带来了前所未有的挑战。文章聚焦硫磺生产线中自动引导车(AGV)系统的设计与优化,旨在克服粉尘对AGV运行的负面影响,提出了一套基于超宽带(UWB)技术和多传感器融合的定位与避障解决方案。方案优化了AGV的路径规划和智能避障机制,并通过集成的粉尘监测传感器提高了作业安全性。测试结果证明,系统能有效适应硫磺粉尘环境,确保稳定和安全的硫磺运输。

Self-tuning weighted measurement fusion Kalman filter and its convergence

For multisensor systems,when the model parameters and the noise variances are unknown,the consistent fused estimators of the model parameters and noise variances are obtained,based on the system identification algorithm,correlation method and least squares fusion criterion.Substituting these consistent estimators into the optimal weighted measurement fusion Kalman filter,a self-tuning weighted measurement fusion Kalman filter is presented.Using the dynamic error system analysis （DESA） method,the convergence of the self-tuning weighted measurement fusion Kalman filter is proved,i.e.,the self-tuning Kalman filter converges to the corresponding optimal Kalman filter in a realization.Therefore,the self-tuning weighted measurement fusion Kalman filter has asymptotic global optimality.One simulation example for a 4-sensor target tracking system verifies its effectiveness.

Self-tuning Information Fusion Kalman Predictor Weighted by Diagonal Matrices and Its Convergence Analysis

为有未知噪音统计的 multisensor 系统，使用现代时间系列分析方法，基于革新建模的动人的一般水准(麻省)的联机鉴定，并且基于为关联功能的矩阵方程的解决方案，噪音变化的评估者被获得，并且在线性最小的变化下面由斜矩阵加权的最佳的信息熔化标准，一个自我调节的信息熔化 Kalman 预言者被介绍，它认识到自我调节的 dec 基于动态错误系统，一个新集中分析方法为自我调节的 fuser 被介绍。在一条认识的集中的一个新概念被介绍，它是比有概率一的集中弱的。如果 MA 革新模型的参数评价是一致的，那么，自我调节的熔化 Kalman 预言者将在一条认识收敛到最佳的熔化 Kalman 预言者，这严格地被证明，或与概率一，以便它有 asymptotic optimality。它能减少计算负担，并且对实时应用合适。为追踪系统的一个目标的一个模拟例子显示出它的有效性。