Search

Battery/Ultracapacitor (UC) Hybrid Energy Storage Systems (HESS) for Electric Vehicles (EVs) have been frequently proposed in the literature to increase battery cycle life. The HESS consists of a Power Management Strategy (PMS) and an Energy Management Strategy (EMS). Existing EMS are quite empirical, such as setting constant target UC energy levels regardless of load. This work presents an improved complete HESS management strategy. The EMS involves a more comprehensive method of setting the target UC energy level using a speed-dependent band. This allows the UC to achieve two goals – contain sufficient energy for future accelerations and have sufficient space for capturing energy from future regenerative braking – without knowledge of the future drive profile. The PMS involves a speed-dependent battery power limit, which also achieves two goals – better UC utilization and allowing the battery to supply the steady state power. Simulations show existing works cannot achieve the four goals simultaneously unless their UCs are sized twice as large compared to the proposed rule-based HESS. In addition, the proposed HESS extends battery cycle life by up to 42% compared to a battery-only system. Lastly, a reduced-scale experiment was built to show that the proposed HESS is able to run in real-time.

Multi-label learning methods can be categorised into algorithm adaptation, problem transformation and ensemble methods. Some of these methods depend on a base classifier and the relationship is not well understood. In this paper the sensitivity of five problem transformation and two ensemble methods to four types of classifiers is studied. Their performance across 11 benchmark datasets is measured using 16 evaluation metrics. The best classifier is shown to depend on the method: Support Vector Machines (SVM) for binary relevance, classifier chains, calibrated label ranking, quick weighted multi-label learning and RAndom k-labELsets; k-Nearest Neighbours (k-NN) and Naive Bayes (NB) for Hierarchy Of Multilabel classifiERs; and Decision Trees (DT) for ensemble of classifier chains. The statistical performance of a classifier is also found to be generally consistent across the metrics for any given method. Overall, DT and SVM have the best performance–computational time trade-off followed by k-NN and NB.

Recently, many methods have been proposed to deal with the problem of cross-domain semantic segmentation. Most of them choose to conduct domain adversarial training either on the high-level convolutional neural network (CNN) features or on the output segmentation maps. Typically, a relatively small weight is given to the adversarial training loss to avoid the problem of mode collapse. However, one potential weakness of these methods is that low-level CNN layers may receive little gradients for domain adaptation, especially when the network is deep. To address this problem, we propose to conduct an auxiliary adversarial training on the fused multi-level CNN features. Gradients for domain adaptation can thus flow into low-level CNN layers more easily along a shorter path. Experiments are conducted on the dataset of Cityscapes with using the source datasets of GTA5 and SYNTHIA, respectively. Quantitative and qualitative results certify the efficacy of the proposed method.

Annotating large-scale datasets to train modern convolutional neural networks is prohibitively expensive and time-consuming for many real tasks. One alternative is to train the model on labeled synthetic datasets and apply it in the real scenes. However, this straightforward method often fails to generalize well mainly due to the domain bias between the synthetic and real datasets. Many unsupervised domain adaptation (UDA) methods were introduced to address this problem but most of them only focused on the simple classification task. This paper presents a novel UDA model which integrates both image and feature level based adaptations to solve the cross-domain object detection problem. We employ objectives of the generative adversarial network and the cycle consistency loss for image translation. Furthermore, region proposal based feature adversarial training and classification are proposed to further minimize the domain shifts and preserve the semantics of the target objects. Extensive experiments are conducted on several different adaptation scenarios, and the results demonstrate the robustness and superiority of the proposed method.

Modern manufacturing enterprises are consuming a considerable amount of energy every year. Improving energy efficiency will not only benefit the enterprises economically, but also help the world to overcome various problems such as energy crisis and air pollution. To achieve this, an accurate energy consumption model is essential. The main objective of this paper is to develop an improved cutting power-based energy consumption model for general end milling process. The proposed model consists of an idle part due to auxiliary components and spindle rotation, and an additional part due to cutting workpiece materials. The first part is modelled as a function of spindle rotation speed, and the other part is considered proportional to the cutting power. Experiments under various milling conditions have demonstrated the effectiveness and efficacy of the proposed model. Comparative studies show that the proposed model is more accurate than other models. Although calibrated from slotting experiments when cutting aluminium alloy, the proposed model is applicable for general milling process. Partial-immersion milling experiments show that the prediction error of the proposed model is as low as 1.74%. When workpiece material changes to titanium alloy, its performance remains decent, with low prediction error of 2.81%. This reveals its capability to provide reliable estimation for different workpiece materials. As such, it could help avoid tedious model calibration, thus saving time, material, and energy. Finally, the energy efficiency of general end milling process is investigated through numerical experiments with the proposed model. By revealing the relationship between energy consumption and various cutting parameters, the proposed model could serve as an excellent platform towards energy-efficient manufacturing/cleaner production.

3D printing and particularly fused deposition modelling (FDM) is widely used for prototyping and fabricating low-cost customised parts. However, present fused deposition modelling 3D printers have limited nozzle condition monitoring techniques to minimize nozzle clogging errors. Nozzle clogging is one of the significant process errors in fused deposition modelling 3D printers, and it affects the quality of prototyped parts in terms of mechanical properties and geometrical accuracy. This paper proposes a dynamic model for current-based nozzle condition monitoring in fused deposition modelling, which is briefly described as follows. First, all the process forces in filament extrusion of the fused deposition modelling were identified and derived theoretically, and theoretical equations of the feed rolling forces and flow-through-nozzle forces were derived. In addition, the effect of the nozzle clogging on the current of extruding motor were identified. Second, based on the proposed dynamic model, current-based nozzle condition monitoring method was proposed. Next, sets of experiments on FDM machine using polylactic acid (PLA) material were carried out to verify the proposed theoretical model, and the results were analysed and evaluated. Findings of the present study indicate that nozzle clogging in FDM 3D printing can be monitored by sensing the current of the filament extruding motor. The proposed model can be used efficiently for monitoring nozzle clogging conditions in fused deposition modelling 3D printers as it is based on the fundamental process modelling.

Background Mesenchymal stem cells (MSCs) can migrate to tissue injury sites where they can induce multipotential differentiation and anti-inflammation effects to treat tissue injury. When traditional therapeutic methods do not work, MSCs are considered to be one of the best candidates for cell therapy. MSCs have been used for treating several injury- and inflammation-associated diseases, including liver cirrhosis. However, the therapeutic effect of MSCs is limited. In some cases, the anti-inflammatory function of naïve MSCs is not enough to rescue tissue injury. Methods Carbon tetrachloride (CCl4) was used to establish a mouse liver cirrhosis model. Enhanced green fluorescence protein (EGFP) and hepatocyte nuclear factor-4α (HNF-4α) overexpression adenoviruses were used to modify MSCs. Three weeks after liver injury induction, mice were injected with bone marrow MSCs via their tail vein. The mice were then sacrificed 3 weeks after MSC injection. Liver injury was evaluated by measuring glutamic-pyruvic transaminase (ALT) and glutamic oxalacetic transaminase (AST) levels. Histological and molecular evaluations were performed to study the mechanisms. Results We found that HNF-4α-overexpressing MSCs had a better treatment effect than unmodified MSCs on liver cirrhosis. In the CCl4-induced mouse liver injury model, we found that HNF-4α-MSCs reduced inflammation in the liver and alleviated liver injury. In addition, we found that HNF-4α promoted the anti-inflammatory effect of MSCs by enhancing nitric oxide synthase (iNOS) expression, which was dependent on the nuclear factor kappa B (NF-κB) signalling pathway. Conclusions MSCs overexpressing HNF-4α exerted good therapeutic effects against mouse liver cirrhosis due to an enhanced anti-inflammatory effect. Gene modification is likely a promising method for improving the effects of cell therapy. Electronic supplementary material The online version of this article (10.1186/s13287-019-1260-7) contains supplementary material, which is available to authorized users.

3D printing and particularly fused filament fabrication is widely used for prototyping and fabricating low-cost customized parts. However, current fused filament fabrication 3D printers have limited nozzle condition monitoring techniques to minimize nozzle clogging errors. Nozzle clogging is one of the most significant process errors in current fused filament fabrication 3D printers, and it affects the quality of the prototyped parts in terms of geometry tolerance, surface roughness, and mechanical properties. This paper proposes a nozzle condition monitoring technique in fused filament fabrication 3D printing using a vibration sensor, which is briefly described as follows. First, a bar mount that supports the liquefier in fused filament fabrication extruder was modeled as a beam excited by a system of process forces. The boundary conditions were identified, and the applied forces were analyzed for Direct and Bowden types of fused filament fabrication extruders. Second, a new 3D printer with a fixed extruder and a moving platform was designed and built for conducting nozzle condition monitoring experiments. Third, nozzle clogging was simulated by reducing the nozzle extrusion temperature, which caused partial solidification of the filament around inner walls of the nozzle. Fourth, sets of experiments were performed by measuring the vibrations of a bar mount during extrusion of polylactic acid, acrylonitrile butadiene styrene, and SemiFlex filaments via Direct and Bowden types of fused filament fabrication extruders. Findings of the current study show that nozzle clogging in fused filament fabrication 3D printers can be monitored using an accelerometer sensor by measuring extruder’s bar mount vibrations. The proposed technique can be used efficiently for monitoring nozzle clogging in fused filament fabrication 3D printers as it is based on the fundamental process modeling.

The resection margin (RM) status and microscopic vascular invasion (MVI) are known prognostic factors for intrahepatic cholangiocarcinoma (ICC). An enhanced understanding of their impact on long-term prognosis is required to improve oncological outcomes.A total of 711 consecutive patients who underwent curative liver resection for hepatitis B virus-related ICC were retrospectively analyzed. The different impact of the RM status (narrow,1 cm, or wide, ≥1 cm) and MVI (positive, +, or negative, -) on overall survival (OS) and recurrence-free survival (RFS) were analyzed.The 1-, 3-, and 5-year OS rates were 67.6%, 42.5%, and 33.2% in wide RMMVI (-), 58.0%, 36.1%, and 26.5% in narrow RMMVI (-), 51.0%, 27.0%, and 24.3% in wide RMMVI (+), and 39.0%, 20.4% and 14.3% in narrow RMMVI (+) (Combined analysis of RM and MVI can better stratify the risks of postoperative death and recurrence in patients with HBV-related ICC, which may help subsequent adjuvant therapy and closer follow-up.

For complex manufacturing process, the raw features are always in a heterogeneous manner that contains lots of high dimensional process parameters and categorical targets (e.g. product quality "Pass" or "Fail") without numerical measurements. It makes the classification problems quite challenging. To improve the classification accuracy, additional data structure is necessarily considered, e.g. feature interactions. Factorization Machines (FM) is a general classification method which considers the effect of feature interactions towards target. This paper discusses the classification performance of FM methods and its extensions to manufacturing processes and other industry fields. Real-world datasets are employed for the validation of classification performance. Experimental results show that FM-based approaches efficiently improve the classification accuracy about 2.49%-5.94% compared with conventional classification methods.

Stress shielding is one of the main problems that lead to bone resorption and revision surgery after implantation. Most of the commercially available metallic non-porous bone implants have a much greater stiffness than the native human bones and are prone to cause stress-shielding. With an open cell structure and intricate architecture, hyperbolic minimal surfaces offer several advantages such as less stress concentration, high permeability and high surface area to volume ratio, thus providing an ideal environment for cell adhesion, migration, and proliferation. This paper explores the use of porous bone implant design based on Triply Periodic Minimal Surfaces (TPMS) which is additively manufactured with ceramic material (Alumina) using Lithography-based Ceramics Manufacturing (LCM) technology. A total of 12 different primitive surface structure unit cells with pore size in the range of 500–1000 μm and porosity above 50% were considered. This is one of the earliest studies reporting the 3D printing of TPMS-based structures using ceramic material. Our results suggest that the choice of material and a porous TPMS-based design led to fabrication of structures with a much lesser compressive modulus comparable with the native bone and hence could potentially be adopted for bone implant design to mitigate the stress-shielding effect.

The design and fabrication of tissue engineering scaffolds is a highly complex process. In order to provide a proper architecture for cells to grow, proliferate, and differentiate to form tissues, scaffolds have to be made with suitable properties. However, the limited structural designs and conventional fabrication techniques severely cripple the improvement of scaffold properties. To overcome these limitations, many researchers have recently adopted computational methods combined with 3D printing techniques as a new approach for scaffold design and fabrication. This approach allows scaffolds to be designed and fabricated with highly complex microstructures and good control and accuracy. Previous works have also shown this approach to be a very useful tool to predict the scaffold properties and to optimize the scaffold designs with a great reduction of experimental iterations. As this approach combining computational methods and 3D printing techniques for scaffold design and fabrication has many advantages over the conventional trial-and-error based approach, it is imperative to provide a state-of-the-art review on the topic. To this end, this article reviews the various applications of computational methods in scaffold design and simulation; it also briefly reviews the application of 3D printing techniques to fabricate the computationally designed scaffolds. Finally, the limitations and future trends of this approach are discussed. Overall, this review will enable readers to understand the benefits of using computational methods coupled with 3D printing to design and fabricate scaffolds, and thus help researchers to improve and optimize the scaffold properties for future tissue engineering research. © 2018 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater 107B: 1329-1351, 2019.

Biomimetic scaffold design is gaining attention in the field of tissue engineering lately. Recently, triply periodic minimal surfaces (TPMSs) have attracted the attention of tissue engineering scientists for fabrication of biomimetic porous scaffolds. TPMS scaffolds offer several advantages, which include a high surface area to volume ratio, less stress concentration, and increased permeability compared to the traditional lattice structures, thereby aiding in better cell adhesion, migration, and proliferation. In literature, several design methods for TPMS scaffolds have been developed, which considered some of the important tissue-specific requirements, such as porosity, Young's modulus, and pore size. However, only one of the requirements of a tissue engineering scaffold was investigated in these studies, and not all of the requirements were satisfied simultaneously. In this work, we develop a design method for TPMS sheet scaffolds, which is able to satisfy multiple requirements including the porosity, Young's modulus, and pore size, based on a parametric optimization approach. Three TPMSs, namely, the primitive (P), gyroid (G), and diamond (D) surfaces, with cubic symmetry are considered. The versatility of the proposed design method is demonstrated by three different applications, namely, tissue-specific scaffolds, scaffolds for stem cell differentiation, and functionally graded scaffolds with biomimetic functional gradients.

Biomimicry is a desirable quality of tissue engineering scaffolds. While most of the scaffolds reported in the literature contain a single pore size or porosity, the native biological tissues such as cartilage and skin have a layered architecture with zone-specific pore size and mechanical properties. Thus, there is a need for functionally graded scaffolds (FGS). EHD-jet 3D printing is a high-resolution process and a variety of polymer solutions can be processed into 3D porous scaffolds at ease, overcoming the limitations of other 3D printing methods (SLS, stereolithography, and FDM) in terms of resolution and limited material choice. In this paper, a novel proof of concept study on fabrication of porous polycaprolactone-based FGS by using EHD-jet 3D printing technology is presented. Organomorphic scaffolds, multiculture systems, interfacial tissue engineering, and in vitro cancer metastasis models are some of the futuristic applications of these polymeric FGS.

Lightweight lattice structures are an important class of cellular structures with high potentials for multi-functional applications. Considering load-bearing requirements, truss buckling is one of the main failure mechanisms for low density and slender lattice structures. Critical buckling loads can be increased by modifying the profile of a truss. In this paper, we present a shape design method to optimize the critical buckling loads for lattice cores with free-form trusses. The free-form truss is represented by Fourier series and implicit surfaces, having smooth truss diameter variations and truss joints. The optimized truss profile is obtained by solving a parametric shape optimization problem with Fourier series coefficients as design variables. The method is used for designing optimized 1D columns and 3D Kagome lattice cores for sandwich panels. The numerical results predict 26.8% and 20.4% improvements of the critical buckling loads for 1D columns and 3D Kagome lattice cores compared to their uniform counterparts of the same mass, respectively. The optimized structures include complex smooth and curved geometries that are well suited for additive manufacturing because of the greater design freedom. Finally, the initial and optimized lattice cores are additively manufactured and tested. The experimental results validate the effectiveness of the proposed method. Keywords: Additive manufacturing, Shape design, Buckling failure, Kagome lattice core, Free-form truss

Object proposal generation, as a preprocessing technique, has been widely used in current object detection pipelines to guide the search of objects and avoid exhaustive sliding window search across images. Current object proposals are mostly based on low-level image cues, such as edges and saliency. However, objectness is possibly a high-level semantic concept showing whether one region contains objects. This paper presents a framework utilizing fully convolutional networks (FCNs) to produce object proposal positions and bounding box location refinement with Support Vector Machine (SVM) to further improve proposal localization. Experiments on the PASCAL VOC 2007 show that using high-level semantic object proposals obtained by FCN, the object recall can be improved. An improvement in detection mean average precision is also seen when using our proposals in the Fast R-convolutional neural network framework. In addition, we also demonstrate that our method shows stronger robustness when introduced to image perturbations, e.g., blurring, JPEG compression, and salt and pepper noise. Finally, the generalization capability of our model (trained on the PASCAL VOC 2007) is evaluated and validated by testing on PASCAL VOC 2012 validation set, ILSVRC 2013 validation set, and MS COCO 2014 validation set.

Part integration is to integrate parts to be a fabrication and assembly unit. It can effectively reduce the fabrication and assembly unit quantity of a product and has been deemed as an effective way to promote the productivity of manufacturing. Although additive manufacturing (AM) has great potential to further promote the part integration for any product (assembly) model, part integration works using AM at present are often ad hoc, human-dependent and time-consuming. One main cause for this problem is that determining which parts in an assembly model can be integrated to be a fabrication and assembly unit automatically is still very difficult, especially when the model has kinematics (inner relative motions embodied by kinematic joints). In this paper, a novel part clustering approach is proposed, based on which, an input assembly model can smartly cluster all its parts to fewer sub-assembly models (each of them fits being integrated to be a fabrication and assembly unit in AM) according to its kinematics. To ensure that the input model after part integration can effectively realize its kinematics using AM, the criteria for part clustering are first defined. Accompanying with the criteria, the methods to determine the kinematics-related fabrication orientation for each part are proposed based on heuristic rules. Then, to make an accurate and efficient part clustering, an attributed part kinematic graph is put forward according to the above criteria. After that, by breaking through the detection automation challenges in sealing support structure and assembly feasibility, an efficient optimization objective function is defined based on the above criteria and graph. Finally, integrating a new adaptive perturbation strategy into the particle swarm optimization algorithm to avoid premature convergence, a novel graph-based part clustering optimization method is designed to cluster all the parts of the input model to be a high-quality (optimized) set of the above-mentioned sub-assembly models. Experiments and analyses are presented to verify the advantages of the proposed approach. Besides, complying with the general guidelines in AM, the proposed approach provides great potential to maximize part integration using AM in a wider application.

Thin-walled tubes have long been used as energy absorbing structures, owing to their high specific energy absorption (SEA) capacity. But, they are associated with substantial high values of peak cr...

Functionally graded lattice core structures show a gradual and localised variation in their mechanical properties with the aim of improving structural performance whilst minimising weight. We present a novel approach to generate optimised functionally graded lattice core structures. Firstly, topology optimisation is performed to return the optimal core density distribution to minimise the structure's compliance subject to a mass constraint. A series of isostatic lines are then constructed with respect to the local principal stresses to generate a lattice structure spatially graded with respect to lattice cell size, aspect ratio and orientation. To validate this novel approach, optimisation is performed on a sandwich core structure subject to three point bending. Experimental tests confirm the greatly improved stiffness and strength properties (101% and 172% respectively) of the core as a result of spatially grading the lattice cells when compared to a benchmark core with uniform cell size of the same density. The new approach also significantly outperforms lattice structures with graded diameters as optimised by state-of-the-art commercial software packages. Non-dimensional core performance indices are formulated to express the relative specific stiffness and strength properties of the core for the three design approaches. Keywords: Topology optimisation, Lattice structures, Finite element analysis, Functionally graded materials, Sandwich structures

3D printing is one of the most innovative technologies in the current era, while 3D bioprinting is revolutionizing the medical technology industry. Bioprinting technology could help overcome the limitations of the current tissue engineering methods, including the problem of longer waiting times for treatment (especially with organ transplants). While fighting infectious diseases had been the main focus of medicine in the past, dealing with the consequences of a predominantly ageing population will be the priority in the future and bioprinting is a promising technology to tackle this challenge effectively. Bioprinting will not only cater the needs of ageing population but also in the field of paediatrics, where the bioprinted tissue or organ should possess the capability to grow with the patient. As researchers around the world are working on 3D bioprinting of tissues and organs, companies are burgeoning all over, making and marketing new bioprinters. While the research and commercialization are moving at such a rapid pace, the issues surrounding the technology, in terms of ethics, policies, regulations and social acceptance, are not addressed in commensurate. Identifying the ELSA (Ethical Legal and Social Aspects) concerns of this technology at an early stage is not only part of our social responsibility but also in the interest of the future of the technology itself. This paper reviews and foresees these challenges with pragmatism, thereby creating awareness to the researchers and policy makers and to urge a positive course of action in the foreseeable future. The significance of this work will be to address a broad audience, associated with this technology, from scientists to businessmen, engineers to clinicians, laymen to lawmakers. A ‘complete’ policy approach for this technology is recommended rather than a ‘piecemeal’ approach of the various constituents of this technology. An effective course of action will be to setup a multi-disciplinary international panel to work on the policy framework, which will look in to both ‘hard’ and ‘soft’ impacts of 3D bioprinting, the associated ethical challenges, legal measures including patenting and effective controls to prevent the misuse, as well as the social aspects encompassing the cultural and religious differences which accounts for the success of this technology. Setting up national level panels to assess the risk-benefit analysis, taking into consideration the cultural and religious view of its population and other legal and social aspects, might be a good starting point.

Modern manufacturing processes are often characterized by high dimensionality and nonlinearity; dimension reduction is usually required to identify the critical features and help improve the process yield. Dimension reduction and analysis of data with high nonlinearity has been a challenging task. Many methods have been proposed based on manifold learning, which aims to learn a lower-dimensional manifold from the original high-dimensional space. However, since the mappings are usually implicit, it is difficult to build a connection between the classification results and the original features. Apart from the prediction results, industry data analysis also focus on the features themselves. Therefore, methods that are able to combine feature selection with dimension reduction is often needed. This paper proposed a hybrid method for nonlinear dimension reduction and feature selection based on Locally Linear Embedding (LLE). LLE and many filter feature selection shares a common procedure of neighbor search. By integrating nearest neighbor search for both LLE and ReliefF, and adjusting the distance measure in supervised LLE with results from feature selection, the proposed method could connect the feature selection process with the nonlinear model, extract the critical features for further analysis, and improve the performance of process modelling. Two dataset from real industry processes are also illustrated as case studies.

Design documents and design project footprints accumulated by corporate information technology systems have increasingly become valuable sources of evidence for design information and knowledge management. Identification and extraction of such embedded information and knowledge into a clear and usable format will greatly accelerate continuous learning from past design efforts for competitive product innovation and efficient design process management in future design projects. Most of the existing design information extraction systems focus on either organizing design documents for efficient retrieval or extracting relevant product information for product optimization. Different from traditional systems, this paper proposes a methodology of learning and extracting useful knowledge using past design project documents from design process perspective based on process mining techniques. Particularly different from conventional techniques that deal with timestamps or event logs only, a new process mining approach that is able to directly process textual data is proposed at the first stage of the proposed methodology. The outcome is a hierarchical process model that reveals the actual design process hidden behind a large amount of design documents and enables the connection of various design information from different perspectives. At the second stage, the discovered process model is analyzed to extract multifaceted knowledge patterns by applying a number of statistical analysis methods. The outcomes range from task dependency study from workflow analysis, identification of irregular task execution from performance analysis, cooperation pattern discovery from social net analysis to evaluation of personal contribution based on role analysis. Relying on the knowledge patterns extracted, lessons and best practices can be uncovered which offer great support to decision makers in managing any future design initiatives. The proposed methodology was tested using an email dataset from a university-hosted multiyear multidisciplinary design project.

Drop-on-demand (DOD) printing is widely used in bioprinting for tissue engineering because of little damage to cell viability and cost-effectiveness. However, satellite droplets may be generated during printing, deviating cells from the desired position and affecting printing position accuracy. Current control on cell injection in DOD printing is primarily based on trial-and-error process, which is time-consuming and inflexible. In this paper, a novel machine learning technology based on Learning-based Cell Injection Control (LCIC) approach is demonstrated for effective DOD printing control while eliminating satellite droplets automatically. The LCIC approach includes a specific computational fluid dynamics (CFD) simulation model of piezoelectric DOD print-head considering inverse piezoelectric effect, which is used instead of repetitive experiments to collect data, and a multilayer perceptron (MLP) network trained by simulation data based on artificial neural network algorithm, using the well-known classification performance of MLP to optimize DOD printing parameters automatically. The test accuracy of the LCIC method was 90%. With the validation of LCIC method by experiments, satellite droplets from piezoelectric DOD printing are reduced significantly, improving the printing efficiency drastically to satisfy requirements of manufacturing precision for printing complex artificial tissues. The LCIC method can be further used to optimize the structure of DOD print-head and cell behaviors.

Soaring energy price, increasing environmental concerns, and stringent legislations have forced enterprises to reduce energy consumption in discrete parts manufacturing. In order to achieve this reduction, accurate estimation of energy consumption at the manufacturing planning stage is needed. In this paper, a new model for energy consumption prediction in machining processes is proposed and validated in experiments. Using the established cutting force model, cutting power at the tool tip is obtained analytically. Further, the relationship between total power consumed by the machine tool and cutting power at the tool tip is empirically characterized. To demonstrate the new model, this hybrid approach has been applied for machine power prediction in a slot milling process. A comparison study has also been conducted between the proposed model and existing models under various cutting conditions. The results show that the proposed model could achieve better prediction accuracy than other models. Moreover, a major limitation of the existing models is successfully addressed. By considering cutting power at the tool tip, the proposed hybrid model could reflect the nature of material removal and is able to provide valuable information regarding the impact of specific cutting parameters on power consumption. This knowledge is important in evaluating operation parameters, process plans, and even production schedule in terms of energy efficiency.

The adaptability of a conventional parser to patent claim parsing is improved.Six claim syntax peculiarities that increase the difficulty of claim parsing are highlighted.New dependency syntax is proposed to handle Long Distance Dependencies.The efficiency of parsing is obviously improved with a segmentation strategy.The improvement on effectiveness of parsing is also promising in a model extraction problem. Patent claim parsing can contribute in many patent-related applications, such as patent search, information extraction, machine translation and summarization. However, patent claim parsing is difficult due to the special structure of patent claims. To overcome this difficulty, the challenges facing the patent claim parsing were first investigated and the peculiarities of claim syntax that obstruct dependency parsing were highlighted. To handle these peculiarities, this study proposes a new two-level parser, in which a conventional parser is imbedded. A patent claim is pre-processed in order to remove peculiarities before passed to the conventional parser. The process is based on a new dependency-based syntax called Independent Claim Segment Dependency Syntax (ICSDS). This two-lever parser has demonstrated promising improvement for patent claim parsing on both effectiveness and efficiency over the conventional parser.

3D bioprinting is a pioneering technology that enables fabrication of biomimetic, multiscale, multi-cellular tissues with highly complex tissue microenvironment, intricate cytoarchitecture, structure-function hierarchy, and tissue-specific compositional and mechanical heterogeneity. Given the huge demand for organ transplantation, coupled with limited organ donors, bioprinting is a potential technology that could solve this crisis of organ shortage by fabrication of fully-functional whole organs. Though organ bioprinting is a far-fetched goal, there has been a considerable and commendable progress in the field of bioprinting that could be used as transplantable tissues in regenerative medicine. This paper presents a first-time review of 3D bioprinting in regenerative medicine, where the current status and contemporary issues of 3D bioprinting pertaining to the eleven organ systems of the human body including skeletal, muscular, nervous, lymphatic, endocrine, reproductive, integumentary, respiratory, digestive, urinary, and circulatory systems were critically reviewed. The implications of 3D bioprinting in drug discovery, development, and delivery systems are also briefly discussed, in terms of in vitro drug testing models, and personalized medicine. While there is a substantial progress in the field of bioprinting in the recent past, there is still a long way to go to fully realize the translational potential of this technology. Computational studies for study of tissue growth or tissue fusion post-printing, improving the scalability of this technology to fabricate human-scale tissues, development of hybrid systems with integration of different bioprinting modalities, formulation of new bioinks with tuneable mechanical and rheological properties, mechanobiological studies on cell-bioink interaction, 4D bioprinting with smart (stimuli-responsive) hydrogels, and addressing the ethical, social, and regulatory issues concerning bioprinting are potential futuristic focus areas that would aid in successful clinical translation of this technology.

Additive manufacturing, commonly referred to as 3D printing, is a technology that builds three-dimensional structures and components layer by layer. Bioprinting is the use of 3D printing technology to fabricate tissue constructs for regenerative medicine from cell-laden bio-inks. 3D printing and bioprinting have huge potential in revolutionizing the field of tissue engineering and regenerative medicine. This paper reviews the application of 3D printing and bioprinting in the field of pediatrics.

With the arrival of cyber physical world and an extensive support of advanced information technology (IT) infrastructure, nowadays it is possible to obtain the footprints of design activities through emails, design journals, change logs, and different forms of social data. In order to manage a more effective design process, it is essential to learn from the past by utilizing these valuable sources and understand, for example, what design tasks are actually carried out, their interactions, and how they impact each other. In this paper, a computational approach based on the deep belief nets (DBN) is proposed to automatically uncover design tasks and quantify their interactions from design document archives. First, a DBN topic model with real-valued units is developed to learn a set of intrinsic topic features from a simple word-frequency-based input representation. The trained DBN model is then utilized to discover design tasks by unfolding hidden units by sets of strongly connected words, followed by estimating the interactions among tasks on the basis of their co-occurrence frequency in a hidden topic space. Finally, the proposed approach is demonstrated through a real-life case study using a design email archive spanning for more than 2 yr.

Mechanical systems with adaptable functions are versatile in functionalities and compact in physical structure than ordinary ones, hence are useful in many industrial and agricultural applications. One aspect of design analysis for these systems is the analysis of functional coupling degree among the adaptable functions, because this is a measurement of how compact and material-effective (thus cost-effective) the systems are. Existing work on functional coupling is mainly oriented to multiple-function systems rather than adaptable-function systems, where the former owes multiple functions without the need of adjustment or replacement of system components. The purpose of this paper is to investigate the methodologies for the evaluation of functional coupling and component importance of adaptable-function systems, so as to provide support for the development of this kind of mechanical systems. To this end, the function-component dependency is investigated first by exploiting the methods for design change propagation, whereby a change possibility matrix is proposed and calculated, and a function-component incidence matrix is developed. Based on the incidence matrix, the functional coupling degree is evaluated to measure the coupling between individual functions, and a comprehensive coupling degree is evaluated to measure the functional coupling of the entire system. These results are then used for the evaluation of component utilization and component importance. Finally, a design case of an adaptable-function juicing machine is studied to illustrate the proposed methods as well as the related concepts.

As closed-loop production becomes an increasingly viable strategy for product life cycle management, cost modelling and analytical tools for closed-loop production systems will become increasingly important. Existing tools based on life cycle costing (LCC) methods are lacking an integrated life cycle perspective in their approach – a perspective that cannot be overlooked when modelling closed-loop production systems. In this paper, we aim to solve this lack of integrated life cycle approach by proposing the product structure-based integrated life cycle analysis (PSILA) technique. It allows us to streamline the process of modelling a closed-loop production system for a complex product based on its product structure. Through this technique, we developed a cost model of a closed-loop production system for a flat-panel (FPD) monitor case study. In this case study, the results simulated by the cost model show an increase in the economic performance when closed-loop production is adopted for the FPD monitor. Deeper analysis of the results in the case study indicate that in order to improve the economic performance of the closed-loop production system for the FPD monitor, the best strategy is not to merely focus on the reutilisation of materials. Instead, the option to remanufacture the FPD monitor for the secondary market should be explored to maximise the value reclaimed from the monitors in the EoL phase. The case study demonstrates that the cost model developed through the PSILA technique has the ability to function as a tool for analysing the economic performance of closed-loop production systems. More importantly, it has the ability to provide us with strategic insights on the cost feasibility of adopting closed-loop production strategies.

Sustainable product development has become an important consideration by many product manufacturers in recent years. Manufacturers are adopting the life cycle approach in the development of products by incorporating the End-of-Life stage considerations at the design stage. To achieve this, however, there is the need to fill up the knowledge gap between the End-of-Life stage and design stage. Tools and methodologies for Design for End-of Life are required to assist designers to design products with better End-of-Life performance. In this paper, a novel Design for End-of-Life methodology is proposed. The methodology captures, represents and analyses the knowledge from the End-of-Life stage into a product End-of-Life Index. The index enables designers to make informed decision on design alternatives for optimal product End-of-Life performance using information from the End-of-Life stage. A case study illustrating the application of this End-of-Life Index on a power tool is subsequently presented. From the discussion, it was shown that the index was able to help to assess the End-of-Life performance of the design and also the changes made subsequently. In essence, this work done here has been demonstrated to be of assistance to designer to adopt Design for End-of-Life.

This paper proposes concept to decide end-of-life (EoL) product recovery option, followed by methods to quantify product condition. A case study is presented using refrigerator crankshaft to illustrate the implication of product condition on product recovery decision making. The product condition comprises wear-out life of the product, change of dimension and cleanliness level. The advantage from the proposed concept is twofold - firstly, knowledge learned and embedded resources from EoL products able to get back to the product life cycle chain as a closed loop and, secondly, the most favourable EoL product recovery option can be made wisely. With thorough understanding of EoL product condition, it enables original equipment manufacturers (OEMs) to make quick and informed decision.

A 3D dental biometrics framework and a pose invariant dental identification (PIDI) technique are proposed for human identification in this study. As best as we can tell, this study is the first attempt at 3D dental biometrics. Using 3D overcomes a number of key hurdles that plague 2D methods in dental identification. 60 Postmortem (PM) samples and 200 Ante mortem (AM) samples taken from multi ethnic Asian groups (Chinese, Indian and Malay) are used in this study. The PIDI technique includes algorithms for feature extraction, feature description and correspondence. The proposed feature extraction algorithm can extract the salient points from the scanned model of dental cast. The proposed feature description and the correspondence algorithm have been tested and shown to be more robust to rigid transformations compared with the related work. Preliminary experimental result achieves 94% rank-1 accuracy in a human-assisted process, while in an automated identification process, the rank-1 accuracy decreases to 80%. In addition, the developed methodology, as it is also feasible to be applied to identifying severely corrupted dental, could promptly provide a potential candidate list in mass disasters before expert investigation. The high accuracy, fast retrieval speed and the facilitated identification process suggest that the developed 3D framework is more suitable for practical use in dental biometrics applications in the future. The limitations and future work are also presented. It could be used adjunctively with the traditional 2D X-ray radiograph identification scheme to improve the efficiency of current identification process.

This paper investigates the feature sampling strategies for 3D partial shape retrieval using bag-of-words model. The SHREC 09’ parts query models [3] are tested for comparison. These parts models are obtained by cutting parts from complete models, which are different from range scans. Dense sampling and pyramid sampling are proposed to extract local salient features from the depth images of the 3D models. Bag-of-words model is used to represent of both of parts query and complete target models. The optimal sampling configurations for the proposed feature extraction strategies are obtained by comparing the retrieval accuracy using maximum histogram intersection distance (MHID). The results suggest that extracting more features does not guarantee better retrieval accuracy using the bag-of-words model. The feature sampling configurations also have significant impacts on the retrieval accuracy.

It is widely agreed that environmental issues, from climate change to energy and water scarcity, to over exploitation of natural resources, represent critical challenges for the society. Thereby, industry has progressively adopted lean manufacturing by reducing amount of energy and materials used in the production. However, manufacturer has limitation towards improving product efficiency without feedback from user as a close loop. Knowing the complexity of environmental issues and economy changes, this paper introduces a new system framework that takes the holistic approach by capturing information at end-of-life (EoL) and feeding back to design or manufacturing in making a greener product. With complete understanding of EoL product status, it enables OEMs to develop technology diversely and at the same time rapidly incorporating environmental concept in part of the product chain. For instance, the valuable knowledge learned in EoL product put forward a more efficient way for product maintenance by cutting down time, cost and negative impact to the environment. Moreover, a greater reliable product could be achieved by utilizing appropriate material or design in particular product. This resulted in improved product design as an inherent loop in product life cycle, therefore producing greener product for better reusability, remanufacturability and recyclability. Overall, efficient EoL product recovery system enables sustainable and competitive manufacturing, and thus drives the economy growth.

To ensure that the economic and environmental impact over the entire product life cycle is taken into consideration during the design process, an integrated life cycle approach to Design for Environment (DfE) must be taken. In this paper, we describe the application of the product structure-based methodology for integrated life cycle analysis as a DfE decision-support tool. While maintaining an integrated perspective of the product life cycle, this methodology modularises the product life cycle model to the individual parts of the product structure; thus, allowing product designers to analyse the parts independently from the rest of the product system and enabling them to systematically evaluate the design alternatives. Using a case study on a dive torchlight, we demonstrate how through our proposed methodology, design alternatives for the product can be evaluated based on the life cycle eco-efficiency.

This paper presents a discretised musculoskeletal multi-body spine model using the LifeMOD Biomechanics Modeller. This was obtained by refining spine segments in cervical, thoracic and lumbar regions into individual vertebra segments, using rotational joints representing the intervertebral discs, building various ligaments between vertebrae and implementing necessary lumbar muscles. To validate the model, two comparison studies were made with in vivo intradiscal pressure measurements of the L4-L5 disc as well as extension moments, axial force and shear force around L5-S1 obtained from spine models available in the literature. The results indicated that the present model is in good correlation with both cases and matches well with experimental data which found that the axial forces are in the range of 3929-4688 N and shear forces up to 650 N. This study provides a preliminary overview of our ongoing work towards building bio-fidelity discretised multi-body spine models for investigating various medical applications. These models can be useful for incorporation into design tools for wheelchairs or other seating systems which may require attention to ergonomics as well as assessing biomechanical behaviour between natural spines and spinal arthroplasty or spinal arthrodesis. Furthermore, these models can be combined with haptic-integrated graphic environments to help surgeons to examine kinematic behaviours of scoliotic spines and to propose possible surgical plans before spine correction operations.

The haptic playback system is important to the training of sensorimotor skills. With the intent for training medical students in surgery, a real-time incision training system with haptic playback is developed in this paper. This system has three modes. Mode I sets the standard cutting trajectory of the incision operation carried out by an expert. Mode II allows students to track the trajectory by haptic guidance to feel the expert's motion. Mode III lets students experience the force feeling of the expert by following the trajectory via the visual cue. To accomplish these tasks, an adjustment proportional-derivative (PD) controller and a cutting model with real-time haptic feedback are devised. The cutting model adopts the mesh adaption method for real-time simulations of progressive cutting operations. These devised methods have been configured with a haptic device PHANToM Desktop to validate the capability of the proposed training system.

Blasting and painting operations involve high consumption of materials such as blasting grits and paint. In Singapore, most of the ship hull cleaning and blasting process are manual operations in the shipyard environment. The shipyard workers need to use fork lifts or cherry pickers, which cause several problems such as low efficiency, harmful pollution to operators' health, and inconsistent blasting quality. In order to improve blasting efficiency, this paper proposes a new design for enclosed blasting chamber mechanism to mimic the manual movements of the blasting guns. To replace the manual blasting operations and increase the efficiency, this mechanism realizes automatic motion control of blasting guns and integrates three blasting guns. A prototype blasting chamber is fabricated for testing experiments. This research is jointly conducted with our industry partner.

In complex manufacturing processes, large amounts of process parameters are monitored and recorded, creating a high-dimensional and heterogonous data warehouse. In order to improve process yield and ensure product quality, comprehensive knowledge of the process should be acquired and critical features should be identified. However, in modern industry production, big data has become quite common; online monitoring and prediction is also usually required. Therefore, an effective feature selection algorithm for industry applications should be fast and robust. Traditional feature selection methods often fails to deal with such demands very well. In this paper, a modified approach for feature selection based on ReliefF is proposed for modelling and analysis of complex manufacturing processes, with improved speed and stability. Randomized k-d tree search is introduced to speed up the feature selection algorithm. The proposed method is also tested with two datasets from real industry process.

Surface cleaning and blasting for the ship hull of oil tankers and passenger ships are conventional operations in a ship yard with surface roughness requirements. Several process parameters affect the blasting quality outcome, such as the distance between the blasting nozzles and ship hull, feed rate of copper grit, grit size, etc. In this paper, a mathematical model is derived to describe the relationship between several input parameters and blasting quality. In addition, a blasting experiment is designed using the Taguchi method in order to reduce the number of experiments required for validating the proposed model. Due to resource and time constraints, the experiments will be carried out as future work.

The increasing design documents created in the design process provide a useful source of process-oriented design information. Hence, the need for automated design information extraction using advanced text mining techniques is increasing. However, most of the existing text mining approaches have problems in mining design information in depth, which results in low efficiency in applying the discovered information to improve the design project. With the aim of extracting process-oriented design information from design documents in depth, this paper proposes a layered text mining approach that produces a hierarchical process model which captures the process behavior at the different level of details. Our approach consists of several interrelated algorithms, namely, a content-based document clustering algorithm, a hybrid named entity recognition (NER) algorithm and a frequency-based entity relationship detection method, which have been integrated into a system architecture for extracting design information from coarse-grained views to fine-grained specifications. To evaluate the performance of the proposed algorithms, experiments were conducted on an email archive that was collected from a real-life design project. The results showed an increase in the detection accuracy for the process-oriented information detection.

Battery/Ultracapacitor (UC) Hybrid Energy Storage Systems for Electric Vehicles have been proposed to increase battery cycle life and driving range. Existing works focus on power management strategies of the battery and UC, while energy management strategies are still relatively empirical. We propose a more rigorous method of setting the target UC energy level using a speed-dependent band. This allows the UC to contain sufficient energy for future accelerations, and have sufficient capacity to store energy from future regenerative braking. Furthermore, by adjusting the band height, the UC size in terms of energy stored can be reduced significantly. We show other works cannot achieve both goals simultaneously unless their UCs are sized larger (up to three times).

Digital prototypes (DPs) allow designers to communicate design concepts to users by rendering physical characteristics of the concepts. To enhance user understanding of the concepts, it is important that the users be able to make better estimates of the values of the characteristics. Correctness of the estimates can depend on two crucial factors: the ability of DPs to render the physical characteristics and the way the DPs are used to communicate the physical characteristics. However, little attention has been paid to the latter. A DP can be used in different ways, e.g., it can be projected on different backgrounds, or be manipulated using different input/output devices. Hence, it is important to identify an effective way of using DPs, via an assessment of the effectiveness of various ways. This paper introduces a methodology for evaluating the effectiveness of communication of physical characteristics to users using DPs. The methodology is used to assess the degree to which users can correctly and quickly estimate the values of the characteristics through interactions with DPs. Such assessments are then analyzed with statistical methods and hypothesis tests to reveal the effectiveness. To validate the proposed methodology, the size of hand-held electronic consumer products, such as smartphones, is considered in a case study. In the study, the effectiveness of two communication setups is evaluated. The same DP is used in both setups, while the environments and input devices are different. The experimental results show that the evaluated effectiveness can reflect how successful the setups are, and can help select the best way of using the DP, i.e., by providing a better environment, a better input device, or a combination of both.

Object proposal is essential for current state-of-the-art object detection pipelines. However, the existing proposal methods generally fail in producing results with satisfying localization accuracy. The case is even worse for small objects which however are quite common in practice. In this paper we propose a novel Scale-aware Pixel-wise Object Proposal (SPOP) network to tackle the challenges. The SPOP network can generate proposals with high recall rate and average best overlap (ABO), even for small objects. In particular, in order to improve the localization accuracy, a fully convolutional network is employed which predicts locations of object proposals for each pixel. The produced ensemble of pixel-wise object proposals enhances the chance of hitting the object significantly without incurring heavy extra computational cost. To solve the challenge of localizing objects at small scale, two localization networks which are specialized for localizing objects with different scales are introduced, following the divide-and-conquer philosophy. Location outputs of these two networks are then adaptively combined to generate the final proposals by a large-/small-size weighting network. Extensive evaluations on PASCAL VOC 2007 show the SPOP network is superior over the state-of-the-art models. The high-quality proposals from SPOP network also significantly improve the mean average precision (mAP) of object detection with Fast-RCNN framework. Finally, the SPOP network (trained on PASCAL VOC) shows great generalization performance when testing it on ILSVRC 2013 validation set., accepted by IEEE Transactions on Image Processing

Vision-based on-road vehicle detection is one of the key problems for autonomous vehicles. Conventional vision-based on-road vehicle detection methods mainly rely on hand-crafted features, such as SIFT and HOG. These hand-crafted features normally require expensive human labor and expert knowledge. Also, they suffer from poor generalization and slow running speed. Therefore, they are difficult to be applied in realistic application which demands accurate and fast detection in all kinds of unpredictable complex environmental conditions. This paper presents a framework utilizing fully convolutional networks (FCN) to produce bounding boxes with high confidence to contain a vehicle, and bounding box location refinement with SVM to further improve localization accuracy. Experiments on the PASCAL VOC 2007 and LISA-Q benchmarks show that using high-level semantic vehicle confidence obtained by FCN, higher precision and recall are achieved. Additionally, FCN enables whole image inference, which makes the proposed method much faster than the object proposal or hand-crafted feature based detectors.

Based on compute analysis engineering (CAE), the performances of acoustic and fluent for muffler comprehensively were analyzed and optimized in the paper. First, aiming at the heavy exhaust noise of diesel engine, bench experiments of measuring independently were designed and carried out, the frequency spectrum of exhaust noise was achieved, combining with simulation analysis of the finite element, acoustic and fluid performances of the original muffler and respective muffling cavities were analyzed, whose shortage such as transmission loss, pressure loss and flow regeneration noise were achieved, the muffler would be re-designed by CAE. Simulation shows the performances of improved mufflers were better than the original one. Eventually, insertion loss of experiment for new muffler is more 10dB(A) above than the original one, and power loss met requirement.