Bayesian Fusion Research Articles

Characterizing the spatial variability of marine soil properties with site-specific sparse data using a Bayesian data fusion approach

Characterizing the spatial variability of marine soil properties with site-specific sparse data using a Bayesian data fusion approach

Imaging research on damage localization of concrete structure based on Lamb wave and Bayesian fusion algorithm

Concrete structures are critical materials in the field of industrial construction and are susceptible to invisible damage under the influence of a variety of factors. The method of increasing the number of sensors is usually used to obtain high-precision damage detection, but it increases the experimental cost and signal processing time, so it is especially important to perform high-precision and low-cost non-destructive testing for it. The article uses Bayes to combine the elliptical trajectory method with a probabilistic damage imaging algorithm, and utilizes the algorithm to achieve higher accuracy damage localization imaging results with a smaller number of sensors. The wavelet transform is used to calculate the time of flight to obtain the localization results of the elliptical trajectory method, the signal damage factor is calculated based on the amplitude information to obtain the localization results of the probabilistic damage imaging, the probability distribution of the damage location is obtained using Bayesian, and the damage localization results are obtained by solving the problem through the Markov chain Monte Carlo method. A dense sensor array based and probabilistic imaging algorithm is also used to localize the damage to the material and the results are compared with the results of the Bayesian fusion algorithm. The results show that the Bayesian fusion algorithm can obtain higher accuracy of damage localization with a significant reduction in the number of sensors, and at the same time greatly reduces the number of signals, simplifies the experimental process, and has high practical value.

Estimation of Vehicular Journey Time Variability by Bayesian Data Fusion With General Mixture Model

Estimation of Vehicular Journey Time Variability by Bayesian Data Fusion With General Mixture Model

Information enhanced slow feature analysis integrated with prior fault data for sensitive monitoring of chemical processes

As an effective feature extraction method, slow feature analysis (SFA) has been successfully applied in the domain of chemical process monitoring. However, as an unsupervised method, the basic SFA ignores the key information carried by prior fault data so that many complicated faults cannot be sensitively detected. To address this issue, an enhanced SFA method, termed Information Enhanced SFA (IE-SFA), is proposed by integrating available fault discriminant information to achieve more sensitive detection of complicated chemical process faults. The proposed method builds a primary-auxiliary SFA modeling framework. On the one hand, the normal training data are analyzed to establish the primary SFA model. On the other, prior fault data are integrated to develop an auxiliary monitoring model for providing extra fault-sensitive information. In the auxiliary model, a sample-variable weighted Fisher discriminant analysis method is performed on normal data, fault data, and their respective slow feature components, thereby extracting fault-sensitive features for enhancing the fault detection capability. For full-view fault monitoring, the Bayesian fusion strategy is used to fuse the outcomes from both primary and auxiliary models. The applications on a benchmark Tennessee Eastman chemical process and a three-phase flow process demonstrate that the proposed IE-SFA method provides superior fault detection performance compared to the basic SFA method.

Output related fault detection and diagnosis based on multi block modified orthogonal Broyden-Fletcher-Goldfarb-Shanno algorithm

In recent years, some progress has been made in fault diagnosis of multi-block regression model, but multi-block method uses the relationship between variables in Hilbert space to block, while the relationship between variables in complex industrial systems often exists in Banach space. In addition, Bayesian fusion method is used in fault monitoring, which can not accurately process the sub-block information and judge whether the fault information is related to the output. In this paper, a new multi-block method is proposed. Firstly, this method uses Ball Covariance (BCov) suitable for Banach space to describe the correlation between variables, and combines K-means clustering method to block variables to ensure that the blocking method is more reasonable. Secondly, an improved orthogonal decomposition Broyden-Fletcher -Goldfarb-Shanno (BFGS) algorithm is proposed. The MOBFGS algorithm performs orthogonal projection on potential structures, removing information orthogonal to the output from the input block. Furthermore, through complete orthogonal decomposition between the input block and the output block, information related to and unrelated to the output is completely separated, and used to detect output related and unrelated faults, respectively. Thirdly, at the block level, MOBFGS algorithm is used for fault detection and contribution graph method for fault diagnosis. Finally, numerical examples and Tennessee Eastman (TE) process simulation verify the effectiveness of this method.

Dempster Combination Rule-Aided Multi-Sensor Decision-Level-Based Data Fusion in Industrial Internet of Things (IIoT) over Zero-Trust Security

Data fusion is a process of accommodating data from discrete sources to generate more relevant information with enhanced accuracy and compatibility rather than the data gathered and information produced by the identical or particular source. To achieve better accuracy in the information inferred by the Industrial Internet of Things (IIoT) system, data fusion is an emerging technology that takes data from different involved devices as input and generates more consistent and accurate information. To safeguard the IIoT data involved in communication foremost is to ensure the authenticity of the devices to block unauthorized access and thus enhance the data confidentiality and data integrity as well. Zero-trust security is thus employed in the IIoT infrastructure to meet the security requirements in question. The proposed approach, Dempster combination rule-aided multi-sensor decision-level-based data fusion (MS-DLDF) in Industrial Internet of Things (IIoT) over zero-trust security, breaks down the whole concept into phases – initially, the device authenticity is ensured before entering and accessing the network, next, the device data is then encoded to avoid data theft or breach from any unauthorized device, and lastly, the data fusion algorithm is executed to gather data from multiple sources as input and reach to more appropriate and relevant information as compared to the information inferred from a single source. The proposed MS-DLDF concludes with the outstanding results of the Dempster combination rule with the increased number of sensors in complex networks as compared to the single-source data fusion and Bayesian estimation data fusion with enhanced accuracy by 6%, minimal computation time by 15%, increased precision rate 6%, reduced false positive rate by 31%, and the amplified data confidentiality rate by 2%, respectively.

Enhancing long-term trend simulation of the global tropospheric hydroxyl (TOH) and its drivers from 2005 to 2019: a synergistic integration of model simulations and satellite observations

Abstract. The tropospheric hydroxyl (TOH) radical is a key player in regulating oxidation of various compounds in Earth's atmosphere. Despite its pivotal role, the spatiotemporal distributions of OH are poorly constrained. Past modeling studies suggest that the main drivers of OH, including NO2, tropospheric ozone (TO3), and H2O(v), have increased TOH globally. However, these findings often offer a global average and may not include more recent changes in diverse compounds emitted on various spatiotemporal scales. Here, we aim to deepen our understanding of global TOH trends for more recent years (2005–2019) at 1×1°. To achieve this, we use satellite observations of HCHO and NO2 to constrain simulated TOH using a technique based on a Bayesian data fusion method, alongside a machine learning module named the Efficient CH4-CO-OH (ECCOH) configuration, which is integrated into NASA's Goddard Earth Observing System (GEOS) global model. This innovative module helps efficiently predict the convoluted response of TOH to its drivers and proxies in a statistical way. Aura Ozone Monitoring Instrument (OMI) NO2 observations suggest that the simulation has high biases for biomass burning activities in Africa and eastern Europe, resulting in a regional overestimation of up to 20 % in TOH. OMI HCHO primarily impacts the oceans, where TOH linearly correlates with this proxy. Five key parameters, i.e., TO3, H2O(v), NO2, HCHO, and stratospheric ozone, can collectively explain 65 % of the variance in TOH trends. The overall trend of TOH influenced by NO2 remains positive, but it varies greatly because of the differences in the signs of anthropogenic emissions. Over the oceans, TOH trends are primarily positive in the Northern Hemisphere, resulting from the upward trends in HCHO, TO3, and H2O(v). Using the present framework, we can tap the power of satellites to quickly gain a deeper understanding of simulated TOH trends and biases.

BayesianSpikeFusion: accelerating spiking neural network inference via Bayesian fusion of early prediction.

Spiking neural networks (SNNs) have garnered significant attention due to their notable energy efficiency. However, conventional SNNs rely on spike firing frequency to encode information, necessitating a fixed sampling time and leaving room for further optimization. This study presents a novel approach to reduce sampling time and conserve energy by extracting early prediction results from the intermediate layer of the network and integrating them with the final layer's predictions in a Bayesian fashion. Experimental evaluations conducted on image classification tasks using MNIST, CIFAR-10, and CIFAR-100 datasets demonstrate the efficacy of our proposed method when applied to VGGNets and ResNets models. Results indicate a substantial energy reduction of 38.8% in VGGNets and 48.0% in ResNets, illustrating the potential for achieving significant efficiency gains in spiking neural networks. These findings contribute to the ongoing research in enhancing the performance of SNNs, facilitating their deployment in resource-constrained environments. Our code is available on GitHub: https://github.com/hanebarla/BayesianSpikeFusion.

Research on dynamic positioning of navigation system using K nearest neighbor algorithm and bayesian fusion based on visible light communication

This study presents a novel Bayesian-K Nearest Neighbor (B-KNN) fusion algorithm aimed at mitigating dynamic positioning errors in navigation systems that utilize visible light communication. The proposed fusion methodology leverages a fingerprint database constructed from the received signal strength data acquired by a mobile receiver. Through the selection of the nearest navigation path points for clustering and the utilization of a Bayesian algorithm with Gaussian fitting, the posterior probability of each cluster is computed to facilitate the prediction of the location of navigation path points via fingerprint matching. The results indicate that 89.8% of simulations exhibit a dynamic positioning error below 1 m within a 20 m×5 m×4 m model, achieving a success rate of 89.8%. Comparative analysis reveals that the B-KNN fusion algorithm enhances dynamic positioning accuracy by 68.3% while reducing computational overhead by approximately 48.3% in contrast to conventional KNN algorithms. Experimental assessments demonstrate that in scenarios involving ambient light interference within a 2 m×1 m×1.5 m model, the majority (95%) of dynamic positioning errors exceed 9 cm, with a significant portion surpassing 10 cm. Conversely, under conditions devoid of ambient light interference, only 4% of errors exceed 9 cm, with 14% falling below 5 cm, indicative of a marked 39.81% improvement in accuracy.

Bayesian multi-exposure image fusion for robust high dynamic range ptychography

The limited dynamic range of the detector can impede coherent diffractive imaging (CDI) schemes from achieving diffraction-limited resolution. To overcome this limitation, a straightforward approach is to utilize high dynamic range (HDR) imaging through multi-exposure image fusion (MEF). This method involves capturing measurements at different exposure times, spanning from under to overexposure and fusing them into a single HDR image. The conventional MEF technique in ptychography typically involves subtracting the background noise, ignoring the saturated pixels and then merging the acquisitions. However, this approach is inadequate under conditions of low signal-to-noise ratio (SNR). Additionally, variations in illumination intensity significantly affect the phase retrieval process. To address these issues, we propose a Bayesian MEF modeling approach based on a modified Poisson distribution that takes the background and saturation into account. The expectation-maximization (EM) algorithm is employed to infer the model parameters. As demonstrated with synthetic and experimental data, our approach outperforms the conventional MEF method, offering superior phase retrieval under challenging experimental conditions. This work underscores the significance of robust multi-exposure image fusion for ptychography, particularly in imaging shot-noise-dominated weakly scattering specimens or in cases where access to HDR detectors with high SNR is limited. Furthermore, the applicability of the Bayesian MEF approach extends beyond CDI to any imaging scheme that requires HDR treatment. Given this versatility, we provide the implementation of our algorithm as a Python package.

Moving Traffic Object Detection Based on Bayesian Theory Fusion

Moving Traffic Object Detection Based on Bayesian Theory Fusion

Mapping reservoir water quality from Sentinel-2 satellite data based on a new approach of weighted averaging: Application of Bayesian maximum entropy

In regions like Oman, which are characterized by aridity, enhancing the water quality discharged from reservoirs poses considerable challenges. This predicament is notably pronounced at Wadi Dayqah Dam (WDD), where meeting the demand for ample, superior water downstream proves to be a formidable task. Thus, accurately estimating and mapping water quality indicators (WQIs) is paramount for sustainable planning of inland in the study area. Since traditional procedures to collect water quality data are time-consuming, labor-intensive, and costly, water resources management has shifted from gathering field measurement data to utilizing remote sensing (RS) data. WDD has been threatened by various driving forces in recent years, such as contamination from different sources, sedimentation, nutrient runoff, salinity intrusion, temperature fluctuations, and microbial contamination. Therefore, this study aimed to retrieve and map WQIs, namely dissolved oxygen (DO) and chlorophyll-a (Chl-a) of the Wadi Dayqah Dam (WDD) reservoir from Sentinel-2 (S2) satellite data using a new procedure of weighted averaging, namely Bayesian Maximum Entropy-based Fusion (BMEF). To do so, the outputs of four Machine Learning (ML) algorithms, namely Multilayer Regression (MLR), Random Forest Regression (RFR), Support Vector Regression (SVRs), and XGBoost, were combined using this approach together, considering uncertainty. Water samples from 254 systematic plots were obtained for temperature (T), electrical conductivity (EC), chlorophyll-a (Chl-a), pH, oxidation–reduction potential (ORP), and dissolved oxygen (DO) in WDD. The findings indicated that, throughout both the training and testing phases, the BMEF model outperformed individual machine learning models. Considering Chl-a, as WQI, and R-squared, as evaluation indices, BMEF outperformed MLR, SVR, RFR, and XGBoost by 6%, 9%, 2%, and 7%, respectively. Furthermore, the results were significantly enhanced when the best combination of various spectral bands was considered to estimate specific WQIs instead of using all S2 bands as input variables of the ML algorithms.

Development of a Bayesian Framework for Kinematic Data Fusion

Structural health monitoring (SHM) is widely used for assessing the condition of bridges at risk. Traditional SHM techniques rely on point-wise information provided by individual sensors placed at strategic locations. However, a more comprehensive assessment of the bridge state can be achieved through data fusion, integrating information from different sensors. This article presents a Bayesian framework data fusion method that combines information from various measurements to improve the knowledge of the structural deformation state. The proposed framework identifies key deformation parameters by exploiting a simplified model that describes the system deformation state and uses an extensive set of data, including prisms, extensometers, tiltmeters, and beyond. Moreover, this approach provides a continuous knowledge of the deformation state, and reduces the uncertainties associated with individual sensor measurements. The framework developed is initially applied to a simulated case study of a simply supported beam, and then to the Colle Isarco viaduct, a highway bridge equipped with an extensive monitoring system.

Bayesian Data Fusion for Enhanced Monitoring of Bridge Displacements Using Satellite InSAR and Topographic Techniques

This paper investigates the effectiveness of satellite Interferometric Synthetic Aperture Radar (InSAR) technology to remotely monitor the structural health of road bridges and detect possible unexpected behaviours. The study is based on an Italian multi-span prestressed concrete bridge and a dataset of eight years of X-band SAR images collected by the COSMO-SkyMed satellite constellation. The method implemented for this study is as follows. First, SAR images are analysed using the Multi-Temporal InSAR technique to extract displacement time series of reflective targets on the bridge decks. Then, the extracted displacements are interpreted to validate their consistency with the bridge's expected responses to environmental loads, such as air temperature and river water flow variations, and detect possible unexpected behaviours. Eventually, the correlation with the air temperature is investigated to identify different bridge spans based only on monitoring data. The findings from this study advance the knowledge on using satellite InSAR-based monitoring for early detection of anomalous bridge behaviours on a large scale, providing valuable insights for the maintenance and safety of critical infrastructure.

Uncertainties quantification for damage localization in concrete based on Bayesian method

The presence of defects in concrete can diminish load-bearing capacity of structures, giving rise to potential concerns regarding safety and durability. Thus, a method that enhances the sensitivity, resolution and robustness of damage localization is critically necessary to assess the condition of concrete structures. This research presents a damage localization method based on Bayesian probabilistic fusion, and uncertainties from measurement and identification process are considered and quantified. The likelihood function is constructed based on the hyperbola-based damage localization method, and the posterior distributions of unknown parameters are calculated via Bayesian theorem combined with measurement data. Furthermore, a meso-level finite element model is established, wherein the concrete medium is considered as a three-phase composite material consisting of polygonal aggregates, mortar matrix and interface transition zones. Owing to the meso-level modeling, the propagation behavior of stress waves within concrete and complicated interactions between stress waves and concrete internal structures can be better characterized. Finally, the damage information, time-difference-of arrival, is extracted from the response signals and the efficiency of the proposed method is verified numerically. The numerical results demonstrate that the proposed probabilistic fusion method outperforms the conventional hyperbola-based method in terms of achieving high spatial resolution and resilience in damage localization.

Robust Detection of Critical Events in the Context of Railway Security Based on Multimodal Sensor Data Fusion.

Effective security surveillance is crucial in the railway sector to prevent security incidents, including vandalism, trespassing, and sabotage. This paper discusses the challenges of maintaining seamless surveillance over extensive railway infrastructure, considering both technological advances and the growing risks posed by terrorist attacks. Based on previous research, this paper discusses the limitations of current surveillance methods, particularly in managing information overload and false alarms that result from integrating multiple sensor technologies. To address these issues, we propose a new fusion model that utilises Probabilistic Occupancy Maps (POMs) and Bayesian fusion techniques. The fusion model is evaluated on a comprehensive dataset comprising three use cases with a total of eight real life critical scenarios. We show that, with this model, the detection accuracy can be increased while simultaneously reducing the false alarms in railway security surveillance systems. This way, our approach aims to enhance situational awareness and reduce false alarms, thereby improving the effectiveness of railway security measures.

Probabilistic remaining useful life prediction without lifetime labels: A Bayesian deep learning and stochastic process fusion method

Trustworthy remaining useful life (RUL) predictions are critical for the long-term safe and reliable operation of degradation systems. The existing deep learning-based methods for RUL prediction are attracting increasing attention but typically face three main challenges. One is the absence of complete run-to-failure data, implying a lack of lifetime labels. Second, it is difficult to directly measure the health indicators (HIs) for field degradation systems. Third, the prediction models output point estimates without uncertainty. To this end, this paper proposes a Bayesian deep learning and stochastic process fusion method for probabilistic RUL prediction without lifetime labels. First, a model-free Bayesian neural network (BNN) is constructed to integrate the quantification of epistemic and aleatoric uncertainties in deep learning. Based on the constructed BNN, it is possible to predict the probability features of HIs. Then, degeneracy modeling is conducted using a nonlinear Wiener process to derive the probability density function of the RUL. Furthermore, model evolution can be achieved through parameter updating during online operations. Finally, the effectiveness and superiority of the proposed prediction method are verified on CALCE battery degradation data.

Investigation into Human Factors in Rear-End Crashes of Commercial Trucks Based on Human Factors Analysis and Classification System with Bayesian Networks Fusion Method

Investigation into Human Factors in Rear-End Crashes of Commercial Trucks Based on Human Factors Analysis and Classification System with Bayesian Networks Fusion Method

An Innovative Correction–Fusion Approach for Multi-Satellite Precipitation Products Conditioned by Gauge Background Fields over the Lancang River Basin

Satellite precipitation products can help improve precipitation estimates where ground-based observations are lacking; however, their relative accuracy and applicability in data-scarce areas remain unclear. Here, we evaluated the accuracy of different satellite precipitation datasets for the Lancang River Basin, Western China, including the Tropical Rainfall Measuring Mission (TRMM) 3B42RT, the Global Precipitation Measurement Integrated Multi-satellitE Retrievals (GPM IMERG), and Fengyun 2G (FY-2G) datasets. The results showed that GPM IMERG and FY-2G are superior to TRMM 3B42RT for meeting local research needs. A subsequent bias correction on these two datasets significantly increased the correlation coefficient and probability of detection of the products and reduced error indices such as the root mean square error and mean absolute error. To further improve data quality, we proposed a novel correction–fusion method based on window sliding data correction and Bayesian data fusion. Specifically, the corrected FY-2G dataset was merged with GPM IMERG Early, Late, and Final Runs. The resulting FY-Early, FY-Late, and FY-Final fusion datasets showed high correlation coefficients, strong detection performances, and few observation errors, thereby effectively extending local precipitation data sources. The results of this study provide a scientific basis for the rational use of satellite precipitation products in data-scarce areas, as well as reliable data support for precipitation forecasting and water resource management in the Lancang River Basin.

Investigating explainable transfer learning for battery lifetime prediction under state transitions

Battery lifetime prediction at early cycles is crucial for researchers and manufacturers to examine product quality and promote technology development. Machine learning has been widely utilized to construct data-driven solutions for high-accuracy predictions. However, the internal mechanisms of batteries are sensitive to many factors, such as charging/discharging protocols, manufacturing/storage conditions, and usage patterns. These factors will induce state transitions, thereby decreasing the prediction accuracy of data-driven approaches. Transfer learning is a promising technique that overcomes this difficulty and achieves accurate predictions by jointly utilizing information from various sources. Hence, we develop two transfer learning methods, Bayesian Model Fusion and Weighted Orthogonal Matching Pursuit, to strategically combine prior knowledge with limited information from the target dataset to achieve superior prediction performance. From our results, our transfer learning methods reduce root-mean-squared error by 41% through adapting to the target domain. Furthermore, the transfer learning strategies identify the variations of impactful features across different sets of batteries and therefore disentangle the battery degradation mechanisms and the root cause of state transitions from the perspective of data mining. These findings suggest that the transfer learning strategies proposed in our work are capable of acquiring knowledge across multiple data sources for solving specialized issues.