Search

Image segmentation is an important research in image processing and machine vision in which automated driving can be seen the main application scene of image segmentation algorithms. Due to the many constraints of power supply and communication in in-vehicle systems, the vast majority of current image segmentation algorithms are implemented based on the deep learning model. Despite the ultrahigh segmentation accuracy, the problem of mesh artifacts and segmentation being too severe is obvious, and the high cost, computational, and power consumption devices required are difficult to apply in real-world scenarios. It is the focus of this paper to construct a road scene segmentation model with simple structure and no need of large computing power under the premise of certain accuracy. In this paper, the ESPNet (Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation) model is introduced in detail. On this basis, an improved ESPNet model is proposed based on ESPNet. Firstly, the network structure of the ESPNet model is optimized, and then, the model is optimized by using a small amount of weakly labeled and unlabeled scene sample data. Finally, the new model is applied to video image segmentation based on dash cam. It is verified on Cityscape, PASCAL VOC 2012, and other datasets that the algorithm proposed in this paper is faster, and the amount of parameters required is less than 1% of other algorithms, so it is suitable for mobile terminals.

Image semantic segmentation as a kind of technology has been playing a crucial part in intelligent driving, medical image analysis, video surveillance, and AR. However, since the scene needs to infer more semantics from video and audio clips and the request for real-time performance becomes stricter, whetherthe single-label classification method that was usually used before or the regular manual labeling cannot meet this end. Given the excellent performance of deep learning algorithms in extensive applications, the image semantic segmentation algorithm based on deep learning framework has been brought under the spotlight of development. This paper attempts to improve the ESPNet (Efficient Spatial Pyramid of Dilated Convolutions for Semantic Segmentation) based on the multilabel classification method by the following steps. First, the standard convolution is replaced by applying Receptive Field in Deep Convolutional Neural Network in the convolution layer, to the extent that every pixel in the covered area would facilitate the ultimate feature response. Second, the ASPP (Atrous Spatial Pyramid Pooling) module is improved based on the atrous convolution, and the DB-ASPP (Delate Batch Normalization-ASPP) is proposed as a way to reducing gridding artifacts due to the multilayer atrous convolution, acquiring multiscale information, and integrating the feature information in relation to the image set. Finally, the proposed model and regular models are subject to extensive tests and comparisons on a plurality of multiple data sets. Results show that the proposed model demonstrates a good accuracy of segmentation, the smallest network parameter at 0.3 M and the fastest speed of segmentation at 25 FPS.

The demand for fluency in human–computer interaction is on an increase globally; thus, the active localization of the speaker by the machine has become a problem worth exploring. Considering that the stability and accuracy of the single-mode localization method are low, while the multi-mode localization method can utilize the redundancy of information to improve accuracy and anti-interference, a speaker localization method based on voice and image multimodal fusion is proposed. First, the voice localization method based on time differences of arrival (TDOA) in a microphone array and the face detection method based on the AdaBoost algorithm are presented herein. Second, a multimodal fusion method based on spatiotemporal fusion of speech and image is proposed, and it uses a coordinate system converter and frame rate tracker. The proposed method was tested by positioning the speaker stand at 15 different points, and each point was tested 50 times. The experimental results demonstrate that there is a high accuracy when the speaker stands in front of the positioning system within a certain range.

Adapting user interface designs for specific tasks performed by different users is a challenging yet important problem. Automatically adapting visualization designs to users and contexts (e.g., tasks, display devices, environments, etc.) can theoretically improve human–computer interaction to acquire insights from complex datasets. However, effectiveness of any specific visualization is moderated by individual differences in knowledge, skills, and abilities for different contexts. A modeling framework called P ersonalized R ecommender System for I nformation visualization M ethods via E xtended matrix completion (PRIME) is proposed for recommending the optimal visualization designs for individual users in different contexts. PRIME quantitatively models covariates (e.g., psychological and behavioral measurements) to predict recommendation scores (e.g., perceived complexity, mental workload, etc.) for users to adapt the visualization specific to the context. An evaluation study was conducted and showed that PRIME can achieve satisfactory recommendation accuracy for adapting visualization, even when there are limited historical data. PRIME can make accurate recommendations even for new users or new tasks based on historical wearable sensor signals and recommendation scores. This capability contributes to designing a new generation of visualization systems that will adapt to users’ states. PRIME can support researchers in reducing the sample size requirements to quantify individual differences, and practitioners in adapting visualizations according to user states and contexts.

A manufacturing system collects big and heterogeneous data for tasks such as product quality modeling and data-driven decision-making. However, as the size of data grows, timely and effective data utilization becomes challenging. We propose an unsupervised data filtering method to reduce manufacturing big data sets with multi-variate continuous variables into informative small data sets. Furthermore, to determine the appropriate proportion of data to be filtered, we propose a filtering information criterion (FIC) to balance the tradeoff between the filtered data size and the information preserved. The case study of a babycare manufacturing and a simulation study have shown the effectiveness of the proposed method.

Additive manufacturing (AM) is a type of advanced manufacturing process that enables fast prototyping to realize personalized products in complex shapes. However, quality defects existed in AM products can directly lead to significant failures (e.g., cracking caused by voids) in practice. Thus, various inspection techniques have been investigated to evaluate the quality of AM products, where X-ray computed tomography (CT) serves as one of the most accurate techniques to detect geometric defects (e.g., voids inside an AM product). Taking a selective laser melting (SLM) process as an example, voids can be detected by investigating CT images after the fabrication of products with limited disturbance from noises. However, limited by the sensor size and scanning speed issue, CT is difficult to be used for online (i.e., layer-wise) voids detection, monitoring, and process control to mitigate the defects. As an alternative, optical cameras can provide layer-wise images to support online voids detection. The intricate texture of the layer-wise image restricts the accuracy of void detection in AM products. Therefore, we propose a new method called pyramid ensemble convolutional neural network (PECNN) to efficiently detect voids and predict the texture of CT images using layer-wise optical images. The proposed PECNN can efficiently extract informative features based on the ensemble of the multiscale feature-maps (i.e., image pyramid) from optical images. Unlike deterministic ensemble strategies, this ensemble strategy is optimized by training a neural network in a data-driven manner to learn the fine-grained information from the extracted feature-maps. The merits of the proposed method are illustrated by both simulations and a real case study in a SLM process.

Additive manufacturing (AM) is considered as a key to personalized product realization as it provides great design flexibility. As the flexibility radically expands the design space, current design space exploration methods for personalized geometric designs become time-consuming due to the use of physically based computer simulations (e.g., finite element analysis or computational fluid dynamics). This poses a significant challenge in design for an efficient personalized product realization cycle, which imposes a tight computation cost constraint to timely respond to every new requirement. To address the challenge, we propose a cost-efficient data-driven design space exploration method for personalized geometric design in AM, enabling feasible design regions under the computation constraint. Specifically, the proposed method adopts surrogate modeling of efficient voxel model-based design rules to identify feasible design regions considering both manufacturability and personalized needs. Since design rules take much less time for evaluation than physically based simulations, the proposed method can contribute to timely providing feasible design regions for an efficient personalized product realization cycle. Moreover, we develop a cost-based experimental design for surrogate modeling, which enables the evaluation of additional design points to provide more precise feasible design regions under the computation cost constraint. The merits of the proposed method are elaborated via additively manufactured microbial fuel cell (MFC) anode design.

In many manufacturing processes, equipment reliability plays a crucial role for product quality assurance. It is important to consider the effect of equipment degradation for the quality-process model. In this article, we propose a dynamic quality-process model to characterize the varying effects of a process to product quality due to equipment degradation. The proposed model considers the effects of process variables on product quality as piecewise linear functions with respect to the equipment degradation. It can automatically estimate the dynamic effects via a meaningful parameter regularization, leading to accurate parameter estimation and model prediction. The merits of the proposed method are illustrated by both simulations and a real case study in a crystal growth manufacturing process.

We propose to use segment graph convolutional and recurrent neural networks (Seg-GCRNs), which use only word embedding and sentence syntactic dependencies, to classify relations from clinical notes without manual feature engineering. In this study, the relations between 2 medical concepts are classified by simultaneously learning representations of text segments in the context of sentence syntactic dependency: preceding, concept1, middle, concept2, and succeeding segments. Seg-GCRN was systematically evaluated on the i2b2/VA relation classification challenge datasets. Experiments show that Seg-GCRN attains state-of-the-art micro-averaged F-measure for all 3 relation categories: 0.692 for classifying medical treatment–problem relations, 0.827 for medical test–problem relations, and 0.741 for medical problem–medical problem relations. Comparison with the previous state-of-the-art segment convolutional neural network (Seg-CNN) suggests that adding syntactic dependency information helps refine medical word embedding and improves concept relation classification without manual feature engineering. Seg-GCRN can be trained efficiently for the i2b2/VA dataset on a GPU platform.

Smart manufacturing, which integrates a multi-sensing system with physical manufacturing processes, has been widely adopted in the industry to support online and real-time decision making to improve manufacturing quality. A multi-sensing system for each specific manufacturing process can efficiently collect the in situ process variables from different sensor modalities to reflect the process variations in real-time. However, in practice, we usually do not have enough budget to equip too many sensors in each manufacturing process due to the cost consideration. Moreover, it is also important to better interpret the relationship between the sensing modalities and the quality variables based on the model. Therefore, it is necessary to model the quality-process relationship by selecting the most relevant sensor modalities with the specific quality measurement from the multi-modal sensing system in smart manufacturing. In this research, we adopted the concept of best subset variable selection and proposed a new model called Multi-mOdal beSt Subset modeling (MOSS). The proposed MOSS can effectively select the important sensor modalities and improve the modeling accuracy in quality-process modeling via functional norms that characterize the overall effects of individual modalities. The significance of sensor modalities can be used to determine the sensor placement strategy in smart manufacturing. Moreover, the selected modalities can better interpret the quality-process model by identifying the most correlated root cause of quality variations. The merits of the proposed model are illustrated by both simulations and a real case study in an additive manufacturing (i.e., fused deposition modeling) process. Published version

Smart manufacturing enables real-time data streaming from interconnected manufacturing processes to improve manufacturing quality, throughput, flexibility, and cost reduction via computation services. In these computation services, machine learning pipelines integrate various types of computation method options to match the contextualized, on-demand computation needs for the maximum prediction accuracy or the best model structure interpretation. On the other hand, there is a pressing need to integrate Fog computing in manufacturing, which will reduce communication time latency and dependency on connections, improve responsiveness and reliability of the computation services, and maintain data privacy. However, there is a knowledge gap in using machine learning pipelines in Fog manufacturing. Existing offloading strategies are not effective, due to the lack of accurate prediction model for the performance of computation services before the execution of those heterogeneous computation tasks. In this paper, machine learning pipelines are implemented in Fog manufacturing. The computation performance of each sub-step of pipelines is predicted and analyzed via linear regression models and random forest regression models. A Fog manufacturing testbed is adopted to validate the performance of the employed models. The results show that the models can adequately predict the performance of computation services, which can be further integrated into Fog manufacturing to better support offloading strategies for machine learning pipelines.

With the vigorous development of artificial intelligence big data era and the advent of 5G era, the amount of network information shows a blowout-like growth. As a result, the accurate query of information is faced with unprecedented challenges. The image as a material reproduction of visual perception, has a large number of retrieval requests all the time, but the traditional image target recognition annotation is mainly based on pixel-level supervised learning. In the face of massive high-quality image recognition, it is very difficult for users to query the target content accurately and quickly. Therefore, this paper studies the animal classification model based on Convolutional Neural Network (CNN), using transfer learning to pre-train the characteristics of the network and combined with the hybrid classification model of CNN. In the experiment, CATS/DOGS were used as the data set, and PyTorch was used to train the network model. Experimental research shows that the accuracy of 96.43% is achieved by using CNN+ transfer learning algorithm, which is significantly higher than that of traditional methods. For small-scale data sets, it effectively solves the non-transferability of manual feature extraction, and improves the accuracy and robustness.

In many science and engineering systems both quantitative and qualitative output observations are collected. If modeled separately the important relationship between the two types of respon...

The objective of this work is to develop a profile-based statistical process control scheme for efficaciously monitoring wafer thickness profiles with non-normality in an industrial wafer slicing process. This is an important research area because the geometric quality of semiconductor wafers in a slicing process has a direct impact on the functional integrity of semiconductor parts and the efficiency of the production. Since non-normality in profiles indicates the existence of inter-cluster variations (i.e., the profiles cannot be represented by a single mean profile with normally-distributed random noise), it deteriorates the effectiveness of many traditional statistical process control methods with normality assumption. To realize the objective of this work, a mixed-effect profile monitoring (MEPM) scheme is proposed. The MEPM scheme adaptively groups profile data into clusters and models the inter-cluster variations, consequently, enabling a robust statistical process control scheme for detecting deviant profile data. Capturing the clustering information of the profile data leads to a deep understanding and an accurate modeling of the spatial data. It is a significant improvement over the current practice of monitoring the geometric product quality by summary quality features (such as total thickness variation) or by profiles neglecting inter-cluster variations. In this paper, the MEPM scheme is tested for detecting the out-of-control wafers from a wafer slicing process. Based on wafer thickness profiles, the MEPM scheme outperforms other benchmark methods and identifies the deviant wafers with low average type II error (missed detection rate) as 0.039. This profile monitoring scheme is extensible to geometrical quality assurance for products in other processes.

The Co2P nanoparticles hybridized with unique N-doping carbon matrices have been successfully designed employing ZIF-67 as the precursor via a facile two-step procedure. The Co2P nanostructures are shielded with reduced graphene oxide (rGO) to enhance electrical conductivity and mitigate volume expansion/shrinkage during sodium storage. As anode materials for sodium-ion batteries (SIBs), the novel architectures of Co2P@N-C@rGO exhibited excellent sodium storage performance with a high reversible capacity of 225 mA h g-1 at 50 mA g-1 after 100 cycles. Our study demonstrates the significant potential of Co2P@N-C@rGO as anode materials for SIBs.

A quantitative understanding of thermal field evolution is vital for quality control in additive manufacturing (AM). Because of the unknown material parameters, high computational costs, and imperfect understanding of the underlying science, physically-based approaches alone are insufficient for component-scale thermal field prediction. Here, we present a new framework that integrates physically-based and data-driven approaches with quasi in situ thermal imaging to address this problem. The framework consists of (i) thermal modeling using 3D finite element analysis (FEA), (ii) surrogate modeling using functional Gaussian process, and (iii) Bayesian calibration based on the thermal imaging data. According to heat transfer laws, we first investigate the transient thermal behavior during AM using 3D FEA. A functional Gaussian process-based surrogate model is then constructed to reduce the computational costs from the high-fidelity, physically-based model. We finally employ a Bayesian calibration method, which compares the surrogate modeling results and thermal measurements, to enable layer-to-layer thermal field prediction across the whole component. A case study on fused deposition modeling is conducted for components with 7 to 16 layers. The cross-validation results show that the proposed framework allows for accurate and fast thermal field prediction for components with different process settings and geometric designs. Keywords: Additive manufacturing, Thermal field, Geometry of freeform, Finite element modeling, Bayesian calibration

Additive manufacturing (AM) enables flexible part geometry and functionality, and reduces product development life cycle by direct layer-wise fabrication from CAD files. In the last decade, great achievements are made on AM materials, machines, processes, etc. However, the quality of the AM parts is still questionable for industrial specifications. On the one hand, AM part quality variables can be either quantitative, such as dimensional accuracy, or qualitative, such as binary indicators for voids, missing features, or surface roughness. On the other hand, both offline process setting variables and functional in situ process variables can be measured and modeled with both quantitative and qualitative (QQ) quality response variables. In this paper, the QQ quality response variables are modeled by offline process setting variables and in situ process variables via functional QQ models. The modeling of these in situ process variables provides the basis for real-time monitoring and control for AM processes. Simulation studies and experimental data from a fused deposition modeling process are performed to demonstrate the effectiveness of the proposed method. Note to Practitioners —Additive manufacturing (AM) processes have attracted much attention and showed many advantages over the traditional subtractive manufacturing processes. However, the product quality issues make AM intractable for high-quality parts in industrial applications. This paper aims to address the quality issues by modeling both quantitative quality variables, such as dimensional accuracy, and qualitative quality variables, such as the binary (go/no-go) indicator for surface conditions. Both offline process setting variables and in situ process variables are used in the model as predictors. Such a model is important for systematically quality evaluation of AM parts, and provides the basis for future process monitoring and control. The merits of the proposed method are demonstrated with simulation studies and a case study in a fused deposition modeling process.

There is a high demand of data visualization providing insights to users in various applications. However, a consistent, online visualization evaluation method to quantify mental workload or user preference is lacking, which leads to an inefficient visualization and user interface design process. Recently, the advancement of interactive and sensing technologies makes the electroencephalogram (EEG) signals, eye movements as well as visualization logs available in user-centered evaluation. This paper proposes a data fusion model and the application procedure for quantitative and online visualization evaluation. 15 participants joined the study based on three different visualization designs. The results provide a regularized regression model which can accurately predict the user's evaluation of task complexity, and indicate the significance of all three types of sensing data sets for visualization evaluation. This model can be widely applied to data visualization evaluation, and other user-centered designs evaluation and data analysis in human factors and ergonomics.

Aerosol jet printing (AJP) is a direct-write-based printed electronics process that is capable to print various materials with fine features [C. Goth et al., “Aerosol jet printing on rapid prototyping materials for fine pitch electronic applications,” in Proc. IEEE 61st Electron. Compon. Technol. Conf., Lake Buena Vista, FL, USA, 2011, pp. 1211-1216.]. It eliminates the complex masking process in traditional semiconductor manufacturing, thus enables flexible electronics designs and reduces manufacturing cost. However, the quality control of AJP processes is still a challenging problem, primarily due to the lack of understanding of the potential root causes of the quality issues. There is a complex interaction among the process setting variables, in situ features, and product quality in AJP processes. In this letter, an ensemble modeling strategy is proposed to quantify the effect of the process setting variables on the in situ features, and the effect of the in situ features on the product quality in a two-level hierarchical way. By identifying significant in situ features as responses for the process setting variables, as well as significant predictors for the product quality in a joint model estimation problem, the proposed model has a hierarchical variable relationship to enable the in situ process control for variation reduction and defect mitigation. A real case study is investigated to demonstrate the advantages of the proposed method.

Graphical models are widely used to model the statistical relationships among variables in a system. Existing graphical models can be used to model the relationships among scalar variables, but cannot be directly applied to model a system with functional variables. In this paper, a novel functional graphical model is proposed to model complex systems where functional variables are measured. To cope with the small sample size problem, we further develop a special sparsity penalization approach to robustly learn the graphical model from limited sample size, and develop a difference from the mean penalization for functional variables. Simulation studies and a case study in a plasma spray manufacturing process are used to demonstrate the effectiveness of the proposed method. Note to Practitioners —Emerging sensing and information technologies have provided unprecedented functional data collection capacity from systems over time. It is important and challenging to model the relationships among these variables. Among many different modeling alternatives, graphical models are powerful tools to characterize the underlying relationships among variables in systems. However, traditional graphical models cannot directly model a system with a mixture of functional and scalar variables. The proposed model aims to address this challenge by proposing a functional graphical model. Based on simulation studies and a case study for a plasma spray manufacturing process, it is shown that the proposed method performs well under various conditions.

Mathematical modeling is an important tool to investigate the performance of microbial fuel cell (MFC) towards its optimized design. To overcome the shortcoming of traditional MFC models, an ensemble model is developed through integrating both engineering model and statistical analytics for the extrapolation scenarios in this study. Such an ensemble model can reduce laboring effort in parameter calibration and require fewer measurement data to achieve comparable accuracy to traditional statistical model under both the normal and extreme operation regions. Based on different weight between current generation and organic removal efficiency, the ensemble model can give recommended input factor settings to achieve the best current generation and organic removal efficiency. The model predicts a set of optimal design factors for the present tubular MFCs including the anode flow rate of 3.47 mL min −1 , organic concentration of 0.71 g L −1 , and catholyte pumping flow rate of 14.74 mL min −1 to achieve the peak current at 39.2 mA. To maintain 100% organic removal efficiency, the anode flow rate and organic concentration should be controlled lower than 1.04 mL min −1 and 0.22 g L −1 , respectively. The developed ensemble model can be potentially modified to model other types of MFCs or bioelectrochemical systems.

In this era of mobile application and socialization, location-based services (LBSs) have become unprecedentedly popular and emphasized. By submitting its location to the service provider, the mobile terminal can obtain useful service. As an important technology that can provide location-based services, spatial query processing has become a research hotspot. According to the problems existing in current road network partitioning, first of all, this paper proposes the partition density formula and the optimal road network partitioning method according to the partition density. Based on that, we propose the concepts of partition self-attraction, mutual attraction, and merger factor of partition to effectively merge the partitions with low densities, which can further reduce the number of unnecessary broadcast frames and optimize the index method. Then, we propose the real distributed air index method DIM and kNN spatial query processing algorithm based on the MapReduce platform. According to the DIM index method, the Name Node only stores the prime index, while not storing the data, and each frame consists of defined index and the data in a partition that need to be broadcasted. The mobile user can quickly localize the frame data that they need to obtain according to the main index or the index of current frame, in this way to reduce the tuning time and access latency, which can significantly optimize the query efficiency. Massive experiments have proved the stability and effectiveness of the proposed algorithm.

Finite element analysis (FEA) has been widely adopted to identify potential defects in additive manufacturing (AM) processes. For personalized product realization, it is necessary to validate a number of heterogeneous product and process designs before or during manufacturing by using FEA. Multi-fidelity FEA simulations can be readily implemented with different capabilities in terms of simulation accuracy. However, due to its complexity, high-fidelity FEA simulation is time-consuming and decreases the efficiency of product realization in AM, while low-fidelity FEA simulation has fast computation speed yet limited capability. Hence, our objective is to improve the capability of FEA by providing an efficient data-driven model. In this research, a Gaussian process-constrained general path model is proposed to approximate the high-fidelity FEA simulation results based on low-fidelity results voxel-by-voxel. The proposed model quantifies the heterogeneous discrepancies between low- and high-fidelity FEA simulation results by incorporating the product design information (e.g., Cartesian coordinates of deposition sequence) and process design information from inputs of FEA simulation (e.g., input heat). Therefore, it enables the validation of new product and process designs based on the simulation results with the desired capability in a timely manner. The advantages of the proposed method are illustrated by FEA simulations of the fused deposition modeling (FDM) process with two levels of fidelity (i.e., low- and high-fidelity).

Lepidolite is a kind of low-grade lithium ore (3.5–4.0% Li2O) which contains more than 22% Al2O3 and 8% K2O. Removal and utilization of the Al and K will be very important for the economic extraction of lithium from this ore. In this paper, a novel method is proposed to remove Al from the leaching solution of lepidolite by adding ammonium from (NH4)2SO4. The results indicate that adding NH4 + could remove more aluminum and have a smaller loss of lithium in comparison to adding K+ for removing aluminum in the leaching solution. The Al removal reached up to 80.44% at (K+ + NH4 +)/Al3+ molar ratio of 0.95, while the amount of Li+ (99.84%) still remained in the leaching solution. In addition, the phase equilibrium in the quaternary K2SO4-Al2(SO4)3-(NH4)2SO4-H2O system at 278.2 K was investigated to support the above conclusion. The results verify that the addition of ammonium could remove more Al3+ by easily forming potassium alum (KAl(SO4)2·12H2O) and ammonium alum (NH4Al(SO4)2·12H2O).

In some manufacturing processes, complex profiles are collected to characterize quality status. However, some of these profiles may have unequal lengths, which makes the attempt of directly comparing them difficult. In addition, when a shift occurs in a profile, it usually affects a segment of continuously connected observations. That is, local shifts instead of global shifts are frequently seen. As shift signals are easily mixed with allowable mean trends, statistical monitoring of such unaligned profiles becomes a challenging task. In this paper, we propose a framework for monitoring profiles with unequal lengths. The profiles are first aligned using a modified robust dynamic time warping algorithm, which is insensitive to local mean shifts. Penalization-based methods are then used to estimate profile means. Finally, mean estimates are utilized in a likelihood ratio test statistic for effective monitoring. Both simulation studies and a real example are used to demonstrate the effectiveness of the proposed monitoring procedure. Copyright © 2016 John Wiley & Sons, Ltd.

In smart manufacturing, data-driven models characterize variable relationships, which are used for decision making to achieve optimal operations. However, data-driven models may not be adequate when modeling assumptions are violated in manufacturing personalization. To advance smart manufacturing to future autonomous manufacturing, we propose data fusion pipelines as a combination of method options (i.e., data fusion or machine learning steps) for manufacturing process modeling. To avoid executing all pipelines, we associate the pipelines with a learning-to-rank method to rank the pipelines with Top-N prediction accuracy, where $N$ is determined by the computation resources. This approach improves the ease of using data fusion and machine learning methods, and effectively avoids large computation workloads in executing of all data fusion pipelines. Case studies in thermal spray coating, aerosol® jet printing, and fused deposition modeling manufacturing are used to demonstrate the effectiveness and efficiency of the proposed approach. The proposed approach is scalable for a larger collection of method options, different manufacturing conditions, and various computation systems and networks.

An industrial cyber-physical system (ICPS) integrates the physical processes, systems and networks with the computation resources to provide reliable and responsive computation services. A cyber-manufacturing system (CMS), which is generated from ICPS, poses significant challenges on reliability, accuracy and responsiveness of computation services for manufacturing decision making. In this paper, we focus on reliability and responsiveness. Due to the heterogeneities of the computation and communication capacities and conditions, demanding computation services may not be completed in a timely manner. To facilitate reliable and responsive computation services, we propose a deadline constrained predictive offloading method based on a mobile-fog-cloud (MFC) network. This method optimizes the offloading decisions by solving a quadratically constrained integer linear programming constrained by latency requirements and predicted availability of devices. A newly constructed hybrid cyber-additive manufacturing network is used to test the performance of the proposed predictive offloading method and the MFC network. The results show that the proposed method outperforms mobile computing, fog computing, cloud computing, and fog-cloud computing benchmarks to minimize resource consumption and comply with latency requirements.

Degradation data have been widely used to estimate product reliability. Because of technology advancement, time-varying usage and environmental variables, which are called dynamic covariates, can be easily recorded nowadays, in addition to the traditional degradation measurements. The use of dynamic covariates is appealing because they have the potential to explain more variability in degradation paths. We propose a class of general path models to incorporate dynamic covariates for modeling of degradation paths. Physically motivated nonlinear functions are used to describe the degradation paths, and random effects are used to describe unit-to-unit variability. The covariate effects are modeled by shape-restricted splines. The estimation of unknown model parameters is challenging because of the involvement of nonlinear relationships, random effects, and shaped-restricted splines. We develop an efficient procedure for parameter estimations. The performance of the proposed method is evaluated by simulations. An outdoor coating weathering dataset is used to illustrate the proposed method. Copyright © 2015 John Wiley & Sons, Ltd.

Aerosol jet printing (AJP) is a direct write technology that enables fabrication of flexible, fine scale printed electronics on conformal substrates. AJP does not require the time consuming mask and postpatterning processes compared with traditional electronics manufacturing techniques. Thus, the cycle time can be dramatically reduced, and highly personalized designs of electronics can be realized. AJP has been successfully applied to a variety of industries, with different combinations of inks and substrates. However, the quality of the printed electronics, such as resistance, is not able to be measured online. On the other hand, the microscopic image sensors are widely used for printed circuit boards (PCBs) quality quantification and inspection. In this paper, two widely used quality variables of printed electronics, resistance and overspray, will be jointly modeled based on microscopic images for fast quality assessment. Augmented quantitative and qualitative (AUGQQ) models are proposed to use features of microscopic images taken at different locations on the printed electronics as input variables, and resistance and overspray as output variables. The association of resistance and overspray can be investigated through the AUGQQ models formulation. A case study for fabricating silver lines with Optomec® aerosol jet system is used to evaluate the model performance. The proposed AUGQQ models can help assess the printed electronics quality and identify important image features in a timely manner.

Single-crystal silicon ingots are produced from a complex crystal growth process. Such a process is sensitive to subtle process condition changes, which may easily become failed and lead to the growth of a polycrystalline ingot instead of the desired monocrystalline ingot. Therefore, it is important to model this polycrystalline defect in the crystal growth process and identify key process variables and their features. However, to model the crystal growth process poses great challenges due to complicated engineering mechanisms and a large amount of functional process variables. In this article, we focus on modeling the relationship between a binary quality indicator for polycrystalline defect and functional process variables. We propose a logistic regression model with hierarchical nonnegative garrote-based variable selection method that can accurately estimate the model, identify key process variables, and capture important features. Simulations and a case study are conducted to illustrate the merits of the proposed method in prediction and variable selection.

Grupiranje (clustering) je jedan od važnih zadataka u rudarenu podataka (data mining), a algoritmi grupiranja utemeljenog na raspodjeli kao što su k-način jedno su od popularnih rješenja. Ipak, sve većim razvojem računarstva u oblaku i ogromne količine podataka, prijenos velikog broja podataka postao je veliki izazov za grupiranje. Na primjer, izvođenje algoritma grupiranja oduzima previše vremena, optimizacija parametara je teška, a kvaliteta grupa (klastera) nije dobra. U tu smo svrhu u ovom radu predložili uobičajeni okvir za algoritme grupiranja utemeljenog na raspodjeli kao što su k-način i dizajnirali njegovu MapReduce implementaciju. Posebice smo, u svrhu predstavljanja prijenosa velikog broja podataka, predložili primjenu tehnike uzorkovanja. Zatim, koristeći k-način algoritam, predlažemo uobičajeni postupak grupiranja i opisujemo primjenu na temelju k-način algoritma. Nadalje, implementiramo predloženi okvir primjenom MapReduce modela programiranja. Eksperimenti pokazuju da je naša metoda učinkovita za prijenos velikog broja podataka., Clustering is one of the significant tasks in data mining, and partition-based clustering algorithms such as k-means are one of the popular solutions. However, with the increasing development of cloud computing and big data, large scale dataset has been a big challenge for clustering. For example, the execution of clustering algorithm is too time-consuming, the optimization of parameters is difficult, and the quality of clusters is not good. To this end, in this paper, we proposed a common framework of partition-based clustering algorithms such as k-means, and designed its MapReduce implementation. Specifically, in order to deal with the representation of large scale dataset, we propose to employ sampling technique. Then, inspired by k-means algorithm, we propose a common procedure of clustering, and provide a k-means based implementation. Furthermore, we implement proposed framework using MapReduce programming model. Experiments show that our method is efficient for large scale dataset.

We are proposing a smart orchestrator as a platform to create and share values in ecosystem in combination with heterogeneous networks. As a smart orchestrator, generative Smart ORCHestrator for converged Tourism ecoSystem (SORCHeTS) supports opportunities to ecologically link multi-organism with symbiotic relationships for the seamlessly plugging. In this research, we are focusing on creation and adaptation of the medical tourism ecosystem on SORCHeTS. The medical tourism as a context-intensive ecosystem is constructed by symbiotic combination of common tourism and medical treatment networks. To do this, it is necessary to combine the keystones and flagships from medical treatment and common tourism organism to create and share productive values. To combine two ecosystems, SORCHeTS are comprised of Genotype and Phenotype modules. Genotype plays a role of creation values in combination with heterogeneous networks. Phenotype dynamically adapts the generative smart ecosystem depending on contextual conditions. Ontology is applied to support dynamic creation and adaptation of the smart ecosystem. To prove effectiveness of the proposed SORCHeTS, we proved interoperability, robustness, creativity, and productivity of newly combined medical tourism ecosystem.

In many manufacturing processes, process data are observed in the form of time-based profiles, which may contain rich information for process monitoring and fault diagnosis. Most approaches currently available in profile monitoring focus on single-stage processes or multistage processes with repeating cyclic profiles. However, a number of manufacturing operations are performed in multiple stages, where non-repeating profiles are generated. For example, in a broaching process, non-repeating cyclic force profiles are generated by the interaction between each cutting tooth and the workpiece. This paper presents a process monitoring method based on Partial Least Squares (PLS) regression models, where PLS regression models are used to characterize the correlation between consecutive stages. Instead of monitoring the non-repeating profiles directly, the residual profiles from the PLS models are monitored. A Group Exponentially Weighted Moving Average (GEWMA) control chart is adopted to detect both global and local shifts. The performance of the proposed method is compared with conventional methods in a simulation study. Finally, a case study of a hexagonal broaching process is used to illustrate the effectiveness of the proposed methodology in process monitoring and fault diagnosis.

Bioelectrochemical systems (BES) are promising technologies to convert organic compounds in wastewater to electrical energy through a series of complex physical-chemical, biological and electrochemical processes. Representative BES such as microbial fuel cells (MFCs) have been studied and advanced for energy recovery. Substantial experimental and modeling efforts have been made for investigating the processes involved in electricity generation toward the improvement of the BES performance for practical applications. However, there are many parameters that will potentially affect these processes, thereby making the optimization of system performance hard to be achieved. Mathematical models, including engineering models and statistical models, are powerful tools to help understand the interactions among the parameters in BES and perform optimization of BES configuration/operation. This review paper aims to introduce and discuss the recent developments of BES modeling from engineering and statistical aspects, including analysis on the model structure, description of application cases and sensitivity analysis of various parameters. It is expected to serves as a compass for integrating the engineering and statistical modeling strategies to improve model accuracy for BES development. Published version