Data Fusion Methodology Research Articles

On-line detection of incipient trend changes in lubricant parameters

Purpose– The purpose of this paper is to present a data fusion methodology for online oil condition and wear particles monitoring for assessment of a mechanical spur gear transmission system.Design/methodology/approach– In this work, a background understanding of the tribological phenomena behind oil degradation and wear on the contact surface of mechanical elements is presented. Experimental results were obtained from oil continuously sampled from an operating a single-stage gearbox. Sampling was done by a multi-sensor automated prototype and online analysis performed by algorithms implemented in a C-code programmed graphical user interface.Findings– Two sets of experiments were performed to observe different fault events frequently occurred in an industrial environment. Fault detection was achieved in appropriate time under constant operating conditions. Under variable operating conditions, same results were obtained by adjusting analysis parameters to critical operation conditions.Originality/value– The value of this research work is the integration of the hardware and software necessary for online monitoring of oil condition and mechanical wear. The setup integrates online sampling with data acquisition, wireless communication, change detection and fault recognition computation. The approach has application in non-destructive online condition-based maintenance.

Matrix weighted multisensor data fusion for INS/GNSS/CNS integration

Inertial navigation system (INS)/global navigation satellite system (GNSS)/ celestial navigation system (CNS) integration is a promising solution to improve the performance of navigation due to the complementary characteristics of INS, GNSS, and CNS. Nevertheless, the information fusion involved in INS/GNSS/CNS integration is still an open issue. This paper presents a matrix weighted multisensor data fusion methodology with two-level structure for INS/GNSS/CNS integrated navigation system. On the first level, GNSS and CNS are integrated with INS by two local filters respectively to obtain local optimal state estimations. On the second level, two different matrix weighted data fusion algorithms, one based on generic weighting matrices and the other based on diagonal weighting matrices, are developed to fuse the local state estimations for generating the global optimal state estimation. These two algorithms are derived in the sense of linear minimum variance, which provide unbiased fusion results no matter whether the local state estimations are mutually independent or not. Thus, they overcome the limitations of the federated Kalman filter by refraining from the use of the upper bound technique. Compared with the data fusion algorithm based on generic weighting matrices, the computational load involved in the one based on diagonal weighting matrices is significantly reduced, even though its accuracy is slightly lower due to the disregard of the coupled relationship between the components of the local state estimations. The effectiveness of the proposed matrix weighted multisensor data fusion methodology is verified through Monte Carlo simulations and practical experiments in comparison with the federated Kalman filter.

The potential relationship discovery model based on result fusion for biomedical medicine research

With the amount of biomedical data growing explosively, medical scientists use many datasets in research for new medicine. However, the amount of biomedical data is growing too fast to abstract hidden information. At the same time, with the development of data storage diversification, scientists prefer to have data fusion based on heterogeneous data sources as opposed to a single data source, and ultimately to achieve knowledge and discovery across heterogeneous databases. Our study focuses on extending the application of latent semantic analysis methodologies into the area of biomedical research. Our purpose is to develop a model for discovering potential relationships between medicines and diseases based on biomedical latent semantic analysis. This model could be used in constructing link maps for biomedical entities, and provide a theoretical basis and practical support for biomedical scientists in their study of the disease–medicine relationship. In detail, we discuss the study of the integration of the latent semantic analysis model and data fusion methodologies. Our result fusion solution combines scientific literature repositories and a biomedical database based on context and the ABC model, and is supervised by a semi-supervised learning algorithm and data fusion algorithms. The expectation is that fused data could represent multilevel potential relationships between biological entities and related emotional relationship expression. The model is validated by experience and proven to be feasible and effective.

Methodologies for Cross-Domain Data Fusion: An Overview

Traditional data mining usually deals with data from a single domain. In the big data era, we face a diversity of datasets from different sources in different domains. These datasets consist of multiple modalities, each of which has a different representation, distribution, scale, and density. How to unlock the power of knowledge from multiple disparate (but potentially connected) datasets is paramount in big data research, essentially distinguishing big data from traditional data mining tasks. This calls for advanced techniques that can fuse knowledge from various datasets organically in a machine learning and data mining task. This paper summarizes the data fusion methodologies, classifying them into three categories: stage-based, feature level-based, and semantic meaning-based data fusion methods. The last category of data fusion methods is further divided into four groups: multi-view learning-based, similarity-based, probabilistic dependency-based, and transfer learning-based methods. These methods focus on knowledge fusion rather than schema mapping and data merging, significantly distinguishing between cross-domain data fusion and traditional data fusion studied in the database community. This paper does not only introduce high-level principles of each category of methods, but also give examples in which these techniques are used to handle real big data problems. In addition, this paper positions existing works in a framework, exploring the relationship and difference between different data fusion methods. This paper will help a wide range of communities find a solution for data fusion in big data projects.

Pinpointing disease genes through phenomic and genomic data fusion.

BackgroundPinpointing genes involved in inherited human diseases remains a great challenge in the post-genomics era. Although approaches have been proposed either based on the guilt-by-association principle or making use of disease phenotype similarities, the low coverage of both diseases and genes in existing methods has been preventing the scan of causative genes for a significant proportion of diseases at the whole-genome level.ResultsTo overcome this limitation, we proposed a rigorous statistical method called pgFusion to prioritize candidate genes by integrating one type of disease phenotype similarity derived from the Unified Medical Language System (UMLS) and seven types of gene functional similarities calculated from gene expression, gene ontology, pathway membership, protein sequence, protein domain, protein-protein interaction and regulation pattern, respectively. Our method covered a total of 7,719 diseases and 20,327 genes, achieving the highest coverage thus far for both diseases and genes. We performed leave-one-out cross-validation experiments to demonstrate the superior performance of our method and applied it to a real exome sequencing dataset of epileptic encephalopathies, showing the capability of this approach in finding causative genes for complex diseases. We further provided the standalone software and online services of pgFusion at http://bioinfo.au.tsinghua.edu.cn/jianglab/pgfusion.ConclusionspgFusion not only provided an effective way for prioritizing candidate genes, but also demonstrated feasible solutions to two fundamental questions in the analysis of big genomic data: the comparability of heterogeneous data and the integration of multiple types of data. Applications of this method in exome or whole genome sequencing studies would accelerate the finding of causative genes for human diseases. Other research fields in genomics could also benefit from the incorporation of our data fusion methodology.

Road Traffic Measurement and Related Data Fusion Methodology for Traffic Estimation

Abstract The knowledge of road traffic parameters is of crucial importance to ensure state-of-the-art traffic services either in public or private transport. In our days, a plethora of road traffic data are continuously collected producing historical and real-time traffic information as well. The available information, however, arrive from inhomogeneous sensor systems. Therefore, a data fusion methodology is proposed based on Switching Kalman Filter. The concept enables efficient travel time estimation for urban road traffic network. On the other hand, the method may contribute to a better macroscopic traffic modelling.

Real-time seam penetration identification in arc welding based on fusion of sound, voltage and spectrum signals

Sensor technology application is the key for intelligent welding process. Multiple sensors fusion has shown their significant advantages over single sensor which can only provide limited information. In this paper, a feature-level data fusion methodology was presented to automatically evaluate seam quality in real time for Al alloy in gas tungsten arc welding by means of online arc sound, voltage and spectrum signals. Based on the developed algorithms in time and frequency domain, multiple feature parameters were successively extracted and selected from sound and voltage signals, while spectrum distribution of argon atoms related to seam penetration were carefully analyzed before feature parameters selection. After the synchronization of heterogeneous feature parameters, the feature-level-based data fusion was conducted by establishing a classifier using support vector machine and 10-fold cross validation. The test results indicate that multisensory-based classifier has higher accuracy i.e., 96.5873 %, than single sensor-based one in term of recognizing seam defects, like under penetration and burn through from normal penetration.

From the Editor-in-Chief - September 2014

In this issue we feature four interesting yet diverse articles related to Aerospace Systems Engineering. The first article describes a new control technique using a 3 Degrees of Freedom (3DOF) model to control a helicopter. The performance of this technique is shown to be superior to three classical control techniques. Next we have two manuscripts both dealing with the subject of sensor information. The first of these discusses techniques for capturing appropriate information from massive amounts of sensor data collected for safety and security and delves into the challenges of data fusion methodologies and impacts this may have on society. The third focuses on multiple sensor tracking architectures and describes one which performs optimally in terms of local track measurements. Our final technical article is one we take great pride in featuring on the cover of this issue to recognize our 2013 Pioneer Award Winners, Prof. Kam Lau and Mr. George Lutes. I think you will find the article very interesting as it describes their work on an ultra-stable fiber-optic system which was implemented and utilized for NASA’s Deep Space Network which in turn provides all of NASA’s interplanetary radio communications. Please join me in congratulating Dr. Lau and Mr. Lutes for their outstanding work which earned them the prestigious Pioneer Award.

Recent Advances on Context-Awareness and Data/Information Fusion in ITS

Intelligent transportation systems (ITS) involve various emerging technologies and applications. This paper presents a comprehensive review of recent advances on data/information fusion and context-awareness referring to ITS. Data/Information fusion is necessary to fuse the data from different sensors and thereby extract relevant information on the target sources. On the other hand, context-aware information processing provides awareness of the driving environments by deploying intelligent query processing and smart information dissemination. The fusion and context-awareness should help in improving ITS operations with better road-awareness service, traffic monitoring, vehicle detection as well as development of new methods. This paper is centered on data fusion and context aware methodologies developed recently in the areas of ITS rather than on their ITS applications. We found that the recent progresses in ITS fusion are devoted to the potential cooperative approaches providing real-time/dynamic vehicle sensing technologies, whereas the recent context awareness techniques are deploying service concepts (e.g. location aware service) and frameworks. It is believed that the newly developed advanced fusion/context-aware techniques are becoming more effective to tackle complex traffic scenarios (e.g. traffic intersection) as well as complex urban environments.

Behavioural data mining of transit smart card data: A data fusion approach

The aim of this study is to develop a data fusion methodology for estimating behavioural attributes of trips using smart card data to observe continuous long-term changes in the attributes of trips. The method is intended to enhance understanding of travellers’ behaviour during monitoring the smart card data. In order to supplement absent behavioural attributes in the smart card data, this study developed a data fusion methodology of smart card data with the person trip survey data with the naïve Bayes probabilistic model. A model for estimating the trip purpose is derived from the person trip survey data. By using the model, trip purposes are estimated as supplementary behavioural attributes of the trips observed in the smart card data. The validation analysis showed that the proposed method successfully estimated the trip purposes in 86.2% of the validation data. The empirical data mining analysis showed that the proposed methodology can be applied to find and interpret the behavioural features observed in the smart card data which had been difficult to obtain from each independent dataset.

Instrumental intelligent test of food sensory quality as mimic of human panel test combining multiple cross-perception sensors and data fusion

Instrumental test of food quality using perception sensors instead of human panel test is attracting massive attention recently. A novel cross-perception multi-sensors data fusion imitating multiple mammal perception was proposed for the instrumental test in this work. First, three mimic sensors of electronic eye, electronic nose and electronic tongue were used in sequence for data acquisition of rice wine samples. Then all data from the three different sensors were preprocessed and merged. Next, three cross-perception variables i.e., color, aroma and taste, were constructed using principal components analysis (PCA) and multiple linear regression (MLR) which were used as the input of models. MLR, back-propagation artificial neural network (BPANN) and support vector machine (SVM) were comparatively used for modeling, and the instrumental test was achieved for the comprehensive quality of samples. Results showed the proposed cross-perception multi-sensors data fusion presented obvious superiority to the traditional data fusion methodologies, also achieved a high correlation coefficient (>90%) with the human panel test results. This work demonstrated that the instrumental test based on the cross-perception multi-sensors data fusion can actually mimic the human test behavior, therefore is of great significance to ensure the quality of products and decrease the loss of the manufacturers.

Drift-Like Terms Minimization in the Responses of a Generic Tin Oxide Gas Sensor

The responses of a tin oxide gas sensor vary with humidity level in the surrounding atmosphere. Such response variations can cause error in the estimation of the concentration level of the target gas and need compensation measures. Different methods have been used to compensate the effect of humidity, which usually require utilization of other parallel environmental sensors and costly data fusion methodology. Particularly, such a drift alters the response patterns obtained from a real or virtual sensor array and hinders gas recognition. Here, we report response patterns recorded from a virtual array made by operating temperature modulation at different ambient humidity levels for three different target gases. Humidity level varied from 30-70% and response patterns were recorded for methanol, ethanol and 1-propanol at a wide concentration range. It is shown that by utilizing the thermal shock-induction method for the temperature modulation of the sensor, the drift levels are low, and with a single set of training data collected at RH=50%, responses obtained in the whole humidity range can be discriminated from each other. The clusters volumes in the feature space grow with the span of the ambient humidity variations, but they remain separate allowing gas recognition.

Medical decision-making inspired from aerospace multisensor data fusion concepts

In recent years, Internet-delivered treatments have been largely used for pain monitoring, offering healthcare professionals and patients the ability to interact anywhere and at any time. Electronic diaries have been increasingly adopted as the preferred methodology to collect data related to pain intensity and symptoms, replacing traditional pen-and-paper diaries. This article presents a multisensor data fusion methodology based on the capabilities provided by aerospace systems to evaluate the effects of electronic and pen-and-paper diaries on pain. We examined English-language studies of randomized controlled trials that use computerized systems and the Internet to collect data about chronic pain complaints. These studies were obtained from three data sources: BioMed Central, PubMed Central and ScienceDirect from the year 2000 until 30 June 2012. Based on comparisons of the reported pain intensity collected during pre- and post-treatment in both the control and intervention groups, the proposed multisensor data fusion model revealed that the benefits of technology and pen-and-paper are qualitatively equivalent . We conclude that the proposed model is suitable, intelligible, easy to implement, time efficient and resource efficient.

Mapping daily evapotranspiration at field scales over rainfed and irrigated agricultural areas using remote sensing data fusion

Continuous monitoring of daily evapotranspiration (ET) at field scale can be achieved by combining thermal infrared remote sensing data information from multiple satellite platforms, given that no single sensor currently exists today with the required spatiotemporal resolution. Here, an integrated approach to field-scale ET mapping is described, combining multi-scale surface energy balance evaluations and a data fusion methodology, namely the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), to optimally exploit spatiotemporal characteristics of image datasets collected by the Landsat and Moderate resolution Imaging Spectroradiometer (MODIS) sensors, as well as geostationary platforms. Performance of this methodology is evaluated over adjacent irrigated and rainfed fields, since mixed conditions are the most challenging for data fusion procedures, and in two different climatic regions: a semi-arid site in Bushland, TX and a temperate site in Mead, NE. Daytime-total ET estimates obtained for the Landsat overpass dates suggest that the intrinsic model accuracy is consistent across the different test sites (and on the order of 0.5mmd−1) when contemporaneous Landsat imagery at 30-m resolution is available. Comparisons between tower observations and daily ET datastreams, reconstructed between overpasses by fusing Landsat and MODIS estimates, provide a means for assessing the strengths and limitations of the fused product. At the Mead site, the model performed similarly for both irrigated and rainfed fields, with an accuracy of about 0.9mmd−1. This similarity in performance is likely due to the relatively large size of the fields (≈50ha), suggesting that the soil moisture dynamics of the irrigated fields are reasonably well captured at the 1-km MODIS thermal pixel scale. On the other hand, the accuracy of daily retrievals for irrigated fields at the Bushland site was lower than that for the rainfed field (errors of 1.5 and 1.0mmd−1, respectively), likely due to the inability of the model to capture ET spikes right after irrigation events for fields substantially smaller than MODIS data resolution. At this site, the irrigated fields were small (≈5ha) compared to the MODIS pixel size, and sparsely distributed over the landscape, so sporadic contributions to ET from soil evaporation due to irrigation were not captured by the 1-km MODIS ET retrievals. However, due the semiarid environment at Bushland, these irrigation-induced spikes in soil evaporation are not long-lived and these underestimations generally affect the irrigation dates only and they do not seem to influence negatively the estimates at the seasonal scale. ET data fusion is expected to perform better over agricultural areas where irrigation is more spatially continuous, resulting in moisture fluxes that are more uniform at the MODIS pixel scale. Overall, the model accurately reproduces the ET temporal dynamics for all the experimental sites, and is able to capture the main differences that were observed between irrigated and rainfed fields at both daily and seasonal time scales.

Ex vivo Time Evolution of Thrombus Growth through Optical and Electrical Impedance data fusion

We designed a novel sensor specifically aimed at ex vivo measurements of white thrombus volume growth; a white thrombus is induced within an artificial micro-channel where hemostasis takes place starting from whole blood under flow conditions. The advantage of the proposed methodology is to identify the time evolution of the thrombus volume by means of an original data fusion methodology based on 2D optical and electrical impedance data simultaneously processed. On the contrary, the present state of the art optical imaging methodologies allow the thrombus volume estimation only at the end of the hemostatic process.

New Developments in Hybrid Optical Coherence Tomographic Imaging: Current Status and Potential Implications in Clinical Practice and Research

Optical coherence tomography (OCT) is a relatively new imaging technique that was introduced to overcome limitations of previous imaging modalities. OCT provides high resolution cross-sectional images of the coronaries which permit detailed evaluation of the luminal morphology and assessment of coronary artery pathology. These unique qualities render OCT a useful tool in clinical practice and research arena. However, apart from its significant advantages OCT imaging has also considerable limitations. To overcome the pitfalls of OCT, fusion of this modality with other imaging techniques has been proposed. Today several hybrid catheter and sophisticated data fusion methodologies have been developed for this purpose, which appear to provide a complete and comprehensive assessment of plaque characteristics and vessel pathophysiology. The aim of this review article is to describe the available OCT based hybrid imaging modalities, present the advantages and limitations of these approaches and discuss their potential value in clinical practice and research arena.

Detecting DDoS Attacks in Cloud Computing Environment

This paper is focused on detecting and analyzing the Distributed Denial of Service (DDoS) attacks in cloud computing environments. This type of attacks is often the source of cloud services disruptions. Our solution is to combine the evidences obtained from Intrusion Detection Systems (IDSs) deployed in the virtual machines (VMs) of the cloud systems with a data fusion methodology in the front-end. Specifically, when the attacks appear, the VM-based IDS will yield alerts, which will be stored into the Mysql database placed within the Cloud Fusion Unit (CFU) of the front-end server. We propose a quantitative solution for analyzing alerts generated by the IDSs, using the Dempster-Shafer theory (DST) operations in 3-valued logic and the fault-tree analysis (FTA) for the mentioned flooding attacks. At the last step, our solution uses the Dempsters combination rule to fuse evidence from multiple independent sources.

Accurate Tree-based Missing Data Imputation and Data Fusion within the Statistical Learning Paradigm

Framework of this paper is statistical data editing, specifically how to edit or impute missing or contradictory data and how to merge two independent data sets presenting some lack of information. Assuming a missing at random mechanism, this paper provides an accurate tree-based methodology for both missing data imputation and data fusion that is justified within the Statistical Learning Theory of Vapnik. It considers both an incremental variable imputation method to improve computational efficiency as well as boosted trees to gain in prediction accuracy with respect to other methods. As a result, the best approximation of the structural risk (also known as irreducible error) is reached, thus reducing at minimum the generalization (or prediction) error of imputation. Moreover, it is distribution free, it holds independently of the underlying probability law generating missing data values. Performance analysis is discussed considering simulation case studies and real world applications.

Application of data fusion methodology for computer forensics dataset analysis to resolve data quality issues in predictive digital evidence

The digital information systems have become increasingly complex and inex- tricably intertwined with the infrastructure of national, public, and private organizations. The forensic digital analysis as a whole, in its relative infancy, is the unwilling victim of the rapid advancement of computer technology, so it is at the mercy of ever more new and complex computing approaches. Forensic digital analysis is unique among the forensic sciences in that it is inherently mathematical and generally comprises more data from an investigation than is present in other types of forensics. The digital investigation process can be driven using numerous forensic investigation models. Among these is the need to analyze forensic materials over complex chains of evidence in a wide variety of hetero- geneous computing platforms. The current computer forensic investigation paradigm is laborious and requires significant expertise on the part of the investigators. This paper presents the application of JDL data fusion model in computer forensics for analyzing the information from seized hard drives along with an analysis of the inter- preted information to prove that the respective user has misused internet. This paper is an attempt to use the data fusion and decision mining processes, to help in enhancing the quality of the investigation process which is in turn is validated by statistical evalua- tion. The mining rules generation process is based on the decision tree as a classification method to study the main attributes that may help in detecting the suspicious behavior. A system that facilitates the use of the generated rules is built which allows investigating agencies to predict the suspicious behavior under study.

A Data Fusion Methodology for Wireless Sensor Systems

An efficient DFA (data fusion algorithm) plays an important role in tracking for moving objects over WSS (wireless sensor system) deployments in order to track the objects accurately. Accuracy in object tracking is mainly dominated by the prediction for those moving targets by filtering and refining the results from wireless mobile sensors deployed in WSS environment. A DFA based on CHHN (competitive Hopfield neural network) technique for obtaining the relationship between measurements results from wireless mobile sensors and estimation of existing tracks over WSS (wireless sensor system) is proposed in this paper. Embedded within the CHNN is also a competitive learning mechanism which creatively removes the dilemma of occasional irrational solutions in traditional HNN (Hopfield neural networks). In this research, except the proposed approach is established with CHNN, the methodology of data fusion over WSS is guaranteed to converge into a stable state when performing a data association. In words, the CHNN-based DFA is combined with wireless mobile sensors in a WSS environment to demonstrate the target tracking capabilities. Computer simulation results illustrate that the new methodology of data fusion based on CHNN is not only successfully able to solve the data association problems addressed over WSS environments, but the specified simulated targets can also be tracked without large scale missing.