An accurate driving cycle prediction is a vital function of an onboard energy management strategy (EMS) for battery/ultracapacitor hybrid storage system (HESS) in electric vehicles. In this paper, we address the requirements to achieve better EMS performances HESS. First, long short-term memory-based method proposed predict cycles under framework model predictive control (MPC) algorithm. Secondly, three EMSs based on fuzzy logic, MPC, and dynamic programming are systematically evaluated...

Brain health attracts more recent attention as the population ages. Smartphone-based gait sensing and analysis can help identify risks of brain diseases in daily life for prevention. Existing approaches mainly hand-craft temporal features or developing CNN-based feature extractors, but they are either prone to lose some inconspicuous pathological information only dedicated a single disease screening. We discover that relationship between segments be used principle generic indicator quantify...

People tracking based on multisensor information fusion is addressed. A framework presented for fusing the laser range finder (LRF) data and monocular camera data. Based this framework, an LRF-based detection algorithm proposed to identify pairs of human legs, by combining motion metric features. Moreover, a geometric observation model developed extract both bearing measurements target person focusing target's shoes with camera. Then, near-optimal particle filter designed fuse from LRF To...

The microstructural deformation-mechanical loading relation of the blood vessel wall is essential for understanding overall mechanical behavior vascular tissue in health and disease. We employed simultaneous loading-imaging to quantify situ deformation individual collagen elastin fibers on unstained fresh porcine coronary adventitia under a combination inflation axial extension loading. Specifically, specimens were imaged biaxial loads study microscopic deformation-loading conjunction with...

The crosslinking reaction process of SPI/EHL-ESO adhesives.

When attempting to replicate the same biological spiking neuron model actions of human brain, methodology and hardware realization design for nervous system brain are crucial considerations. This work provides a modified neural modeling complete Digital Spiking Silicon (DSSN4D). is capable regenerating basic attributes original using simplified power-2 based technique. The suggested on fundamental operations that can be implemented similar main model. Removing nonlinear parts (original one)...

Abstract Introduction While smoking has been associated with many negative consequences to human health, one possible benefit is that nicotine could improve cognitive functions. Previous studies have suggested may influence brain-derived neurotrophic factor (BDNF) levels. However, the exact effects of on BDNF and cognition, potential neurobiological mechanisms underlying changes induced by nicotine, remains elusive. Methods Two tobacco products were used in population detect plasma BDNF,...

With the development of Internet Things (IoT), indoor localization has been a research focus in recent years. For inertial measurement unit (IMU)-based method, zero velocity update (ZUPT) uses known at stationary epoch as benchmark to calibrate drift. However, only takes up 24% whole gait cycle time, and drift remaining 76% time is usually estimated according an assumption that linear over which would introduce errors. In this paper, two-step calibration method was proposed based on human...

Conversational Recommender Systems (CRSs) in E-commerce platforms aim to recommend items users via multiple conversational interactions. Click-through rate (CTR) prediction models are commonly used for ranking candidate items. However, most CRSs suffer from the problem of data scarcity and sparseness. To address this issue, we propose a novel knowledge-enhanced deep cross network (K-DCN), two-step (pretrain fine-tune) CTR model We first construct billion-scale conversation knowledge graph...

Visually-rich Document Understanding (VrDU) has attracted much research attention over the past years.Pre-trained models on a large number of document images with transformer-based backbones have led to significant performance gains in this field.The major challenge is how fusion different modalities (text, layout, and image) documents unified model pre-training tasks. This paper focuses improving text-layout interactions proposes novel multi-modal model, LayoutMask.LayoutMask uses local 1D...

We have combined hydrogen-bonding complexation in solution and layer-by-layer assembly for the controlled loading of a water-insoluble small organic molecule, bis-triazine (DTA), an azobenzene derivative containing multiple hydrogen bond donors acceptors, into multilayer films poly(acrylic acid) diazo-resin. UV-visible spectroscopy indicates that DTA has been loaded films, with amount increasing linearly number layers. The can be well tuned either by changing concentration or solvent...

With the development of technologies, an increasing number wearable devices that are currently at heart Internet Things used around world. The concerns about privacy, in particular healthcare devices, exacerbated as requirement for self-health monitoring increases. In addition, many sources data geographically separated and might not be allowed to release due patients or regulatory constraints. Accordingly, this paper, a distributed hierarchical deep learning system is proposed. proposed...

To develop a deep learning-based method, dubbed Denoising CEST Network (DECENT), to fully exploit the spatiotemporal correlation prior image denoising.DECENT is composed of two parallel pathways with different convolution kernel sizes aiming extract global and spectral features embedded in images. Each pathway consists modified U-Net residual Encoder-Decoder network 3D convolution. Fusion 1 × utilized concatenate pathways, output DECENT noise-reduced The performance was validated numerical...

Recent advances in multimodal pre-trained models have significantly improved information extraction from visually-rich documents (VrDs), which named entity recognition (NER) is treated as a sequence-labeling task of predicting the BIO tags for tokens, following typical setting NLP. However, BIO-tagging scheme relies on correct order model inputs, not guaranteed real-world NER scanned VrDs where text are recognized and arranged by OCR systems. Such reading issue hinders accurate marking...

Phenol-formaldehyde (PF) resin, a significant synthetic thermosetting polymer, is extensively utilized across diverse engineering domains. The exploration of the high-temperature oxidation mechanism PF resin pivotal to enhancing its thermal stability. However, current research lacks comprehensive study on pyrolysis under different oxidizing conditions. Herein, this work systematically explores various environments by using experiments and molecular dynamics simulations. oxidative cracking...

Early diagnosis of depression is crucial for effective treatment. Our study utilizes functional near-infrared spectroscopy (fNIRS) and machine learning to accurately classify mild severe depression, providing an objective auxiliary diagnostic tool mental health workers.

In recent days, the complexity of software has increased significantly in embedded products such a way that verification (ESW) now plays an important role to ensure product's quality. Embedded systems engineers usually face problems verifying properties have meet application's deadline, access memory region, handle concurrency, and control hardware registers. This work proposes semiformal approach combines dynamic static stress cover exhaustively state space system. We perform case study on...