Search

Maximum power point tracking (MPPT) technology is well suited to photovoltaic (PV) system analysis, which can improve power generation efficiency and control effectiveness. However, the nonlinear intrinsic characteristics of the PV system cannot guarantee that the power generation system always works stably at the maximum power point (MPP). Additionally, the accuracy and complexity of the system modeling are crucial to the performance of the PV system. In this article, a novel adaptive Rao-Blackwellized cubature Kalman filter (ARBCKF) scheme is developed that can be efficiently applied to estimate MPPT for PV systems. At first, a comprehensive analysis of the cubature Kalman filter (CKF) is conducted in the context of MPPT. Second, the LogWright function is utilized to represent the current as an explicit function of the voltage, which achieves faster and more accurate execution. Finally, an adaptive-Rao-Blackwellized method is applied to dynamic systems with nonlinear equations of state and linear equations of measurement, which can capitalize on adapting the noise covariance and improving the performance. Compared with the widely adopted other Kalman filter methods, computation results show that MPP estimation accuracy can be significantly increased in diverse transient and steady-state scenarios.

Most existing road extraction methods prioritize region accuracy at the expense of ignoring road boundaries and connectivity quality. The occluding objects such as buildings, trees, and vehicles in remote sensing data usually cause discontinuous mask outputs, and consequently affect road extraction accuracy. In this article, a road extraction fusion network perceiving region and boundary features is proposed. The combination of a location-aware transformer and convolutional neural network is responsible for focusing regional semantic information through adaptive weight filtering. Combining spatial and channel information, the graph convolutional network is improved by constructing an integrated adjacency matrix to consider the relationships between nodes at different scales, which allows for better capture of multiscale contexts. Boundary details are used to complement regional features, thereby enhancing the connectivity of masks. Comprehensive quantitative and qualitative experiments demonstrate that our method significantly outperforms state-of-the-art methods on two public benchmarks, which can improve road extraction by handling interruptions related to shadows and occlusions, producing high-resolution masks.

This study presents a novel massively high-throughput reinforcement learning (RL) framework specifically designed for addressing classic control problems, leveraging our proposed architecture and algorithms optimized for efficient concurrent computations on GPUs. Our research demonstrates the effectiveness of our methods in efficiently training RL agents across various classic control problems, encompassing both discrete and continuous domains, while achieving rapid and stable performance up to 10K concurrent environment instances. Furthermore, we observe that RL exploration with a large number of parallel instances significantly enhances the stability of updating a shared model. For instance, we show that the stability of Deep Deterministic Policy Gradient (DDPG) training can be achieved without requiring experience replay, as evidenced in our study.

Clutter removal in ground-penetrating radar (GPR) based on deep learning has been studied in recent years. However, existing methods are primarily designed for homogeneous background conditions and utilize only local spatial information via the convolution operation. In order to solve these issues, a subspace projection attention (SPA) network is proposed for GPR heterogeneous clutter removal in this article. First, a heterogeneous concrete dataset based on a numerical model with randomly placed aggregates is constructed, which incorporates the complex electromagnetic propagation process accurately to improve the effectiveness for heterogeneous clutter removal. In addition, the clutter basis learning neural network is designed by integrating the SPA module into the skip connection paths of U-Net architecture. By learning the subspace basis vectors adaptively, the SPA exploits both local and global spatial information to extract target features precisely. At the same time, the feature maps are projected to the target subspace to remove heterogeneous clutter features. Finally, the performance and effectiveness of proposed method are validated by simulations and experiments.

Hollow-core negative-curvature fibres have been the focus of current research, but how to reduce the loss of hollow-core negative-curvature fibres is a serious problem. This paper proposes a new structure of hollow-core negative-curvature fibres, that is numerically simulated using the finite-element method, and compared with the simulation results of hollow-core conjoined-tube negative curvature fibres (CTNCFs) and hollow-core nested anti-resonant nodeless fibres (NANFs). The results show that, the LP01 mode loss can reach 0.003694 dB/km with a transmission wavelength of 1.06 μm, the LP11 mode loss can be as low as 0.28423 dB/km, and the bending loss is 0.4405 dB/km with a bending radius of 5 cm and a transmission wavelength of 1.06 μm. The loss is reduced by an order of magnitude compared to CTNCFs and NANFs.

This paper reports a pioneering demonstration platform of a cryogenic propulsion unit at liquid nitrogen temperature. A high temperature superconducting (HTS) machine is connected with a cryogenic power rectifier in a generator mode to prove the feasibility of a cryogenic propulsion unit for electric aircraft propulsion applications. Machine operation was carried out with a special focus on the total heat dissipation inside the HTS AC windings. Different types of 2G HTS tapes were tested to provided data for stator coils’ design. The transient operation was carried out to represent a short-circuit failure in one of the power electronic devices. The test shows AC loss performance of the HTS windings using calorimetric method during the short circuit event, indicating the importance of developing protection schemes for the cryogenic propulsion units to prevent damage to the HTS components. The paper provides initial insights into the interaction between superconducting machines and cryogenic power electronics within a cryogenic propulsion unit. It is an important first step to understand and further develop cryogenic propulsion technology for future electric aircraft.

Computer-aided diagnosis systems with deep learning frameworks have been used to identify benign and malignant pulmonary nodules in lung cancer diagnosis. It's commonly known that a premise of training complex deep neural nets is the large-scale labeled datasets. However, the abundance of labeled datasets is usually unavailable in many medical image domains. This factor can lead to the poor generalization performance of deep learning models. In this paper, we propose a novel multi-discriminator generative adversarial network model combined with an encoder for the classification of benign and malignant pulmonary nodules. To the best of our knowledge, we are the first to apply unsupervised learning to identify benign and malignant lung nodules. Firstly, we use a multi-discriminator generative adversarial network to build a generative model trained with unlabeled benign lung nodule images. Then an encoder is combined with the trained generative model to establish a mapping of benign pulmonary nodule images to the latent space. The benign and malignant lung nodules are scored by calculating the GAN discriminator feature loss and image reconstruction loss. The model yields high anomaly scores on malignant images and low anomaly scores on benign images. Experimental results show that our method with only a small number of unlabeled datasets could achieve more competitive results compared with other supervised deep learning approaches.

The redundant convolutional encoder-decoder network has been proven useful in speech enhancement tasks. This network can capture the localized time-frequency details of speech signals through the fully convolutional network structure and the feature selection capability that results from the encoder-decoder mechanism. However, extracting informative features, which we regard as important for the representational capability of speech enhancement models, is not considered explicitly. To solve this problem, we introduce the attention mechanism into the convolutional encoder-decoder model to explicitly emphasize useful information from three aspects, namely, channel, space, and concurrent space-and-channel. Furthermore, the attention operation is specifically achieved through the squeeze-and-excitation mechanism and its variants. The model can adaptively emphasize valuable information and suppress useless ones by assigning weights from different perspectives according to global information, thereby improving its representational capability. Experimental results show that the proposed attention mechanisms can employ a small fraction of parameters to effectively improve the performance of CNN-based models compared with their normal versions, and generalize well to unseen noises, signal-to-noise ratios (SNR) and speakers. Among these mechanisms, the concurrent space-channel-wise attention exhibits the most significant improvement. And when comparing with the state-of-the-art, they can produce comparable or better results. We also integrate the proposed attention mechanisms with other convolutional neural network (CNN)-based models and gain performance. Moreover, we visualize the enhancement results to show the effect of the attention mechanisms more clearly.

The auditory selection framework with attention and memory (ASAM), which has an attention mechanism, embedding generator, generated embedding array, and life-long memory, is used to deal with mixed speech. When ASAM is applied to speech enhancement, the discrepancy between the voice and noise feature memories is huge and the separability of noise and voice is increased. However, ASAM cannot achieve desirable performance in terms of speech enhancement because it fails to utilize the time-frequency dependence of the embedding vectors to generate a corresponding mask unit. This work proposes a novel embedding encoder-decoder (EED), and a convolutional neural network (CNN) is used as decoder. The CNN structure is good at detecting local patterns, which can be exploited to extract correlation embedding data from the embedding array to generate the target spectrogram. This work evaluates a similar ASAM, EED with an LSTM encoder and a CNN decoder (RC-EED), RC-EED with an attention mechanism (RC-AEED), other similar EED structures and baseline models. Experiment results show that RC-EED and RC-AEED networks have good performance on speech enhancement task at low signal-to-noise ratio conditions. In addition, RC-AEED exhibits superior speech enhancement performance over ASAM and achieves better speech quality than do deep recurrent network and convolutional recurrent network.

Binary code analysis is crucial in various software engineering tasks, such as malware detection, code refactoring, and plagiarism detection. With the rapid growth of software complexity and the increasing number of heterogeneous computing platforms, binary analysis is particularly critical and more important than ever. Traditionally adopted techniques for binary code analysis are facing multiple challenges, such as the need for cross-platform analysis, high scalability and speed, and improved fidelity, to name a few. To meet these challenges, machine learning-based binary code analysis frameworks attract substantial attention due to their automated feature extraction and drastically reduced efforts needed on large-scale programs. In this paper, we provide the taxonomy of machine learning-based binary code analysis, describe the recent advances and key findings on the topic, and discuss the key challenges and opportunities. Finally, we present our thoughts for future directions on this topic.

The phenomenon of efficiency droop is comprehensively investigated in an asymmetric GaN-based laser diode (LD). Numerical simulations and experiments are both conducted. It is found that with the introduction of a sandwiched GaN/InAlN/GaN lower quantum barrier (LQB) instead of the bulk GaN LQB in an asymmetric GaN-based high-power blue LD, the efficiency droop is effectively mitigated to only 9.7% under continuous-wave operation with no degradation in the crystal quality. Compared with the conventional asymmetric GaN-based LD, the wall-plug efficiency is successfully increased by 24.7% from 15.8% to 19.7% at high injection current of 2000 mA, promoting the increase of slope efficiency and output power, while the threshold current is reduced by 15.2%. These improvements are primarily attributed to the inserted high-quality InAlN thin barrier layer, which could form an extra potential barrier beneath the first InGaN quantum well to remarkably alleviate the hole overflow from the active region and enhance the radiative recombination rate.

The indium–tin-oxide-free organic solar cells (OSCs) have been fabricated using a semitransparent Au film as anode. A comparative study between the OSCs with different metal cathodes shows that formation of an ideal ohmic contact at low-workfunction Ca and the organic interface reduces series resistance and increases the carrier collection. Moreover, a periodic corrugation was introduced into the device structure, which leads to an enhanced light absorption in the OSCs by elongating the optical path of the incident light inside the absorber material. The highest power conversion efficiency of an optimal OSC with a Au anode and a corrugated Ca/Ag cathode is 1.4%, which is significantly improved in contrast to 0.61% for the OSC with a planar Al cathode.

Synergetic recovery of both somatosensory and motor functions is highly desired by limb amputees to fully regain their lost limb abilities. The commercially available prostheses can restore the lost motor function in amputees but lack intuitive sensory feedback. The previous studies showed that electrical stimulation on the arm stump would be a promising approach to induce sensory information into the nervous system, enabling the possibility of realizing sensory feedback in limb prostheses. However, there are currently limited studies on the effective evaluation of the sensations evoked by transcutaneous electrical nerve stimulation (TENS). In this paper, a multichannel TENS platform was developed and the different stimulus patterns were designed to evoke stable finger sensations for a transradial amputee. Electroencephalogram (EEG) was recorded simultaneously during TENS on the arm stump, which was utilized to evaluate the evoked sensations. The experimental results revealed that different types of sensations on three phantom fingers could be stably evoked for the amputee by properly selecting TENS patterns. The analysis of the event-related potential (ERP) of EEG recordings further confirmed the evoked sensations, and ERP latencies and curve characteristics for different phantom fingers showed significant differences. This work may provide insight for an in-depth understanding of how somatosensation could be restored in limb amputees and offer technical support for the applications of non-invasive sensory feedback systems. [ABSTRACT FROM AUTHOR]

In this paper, a novel denoising method for electrocardiogram (ECG) signal is proposed to improve performance and availability under multiple noise cases. The method is based on the framework of conditional generative adversarial network (CGAN), and we improved the CGAN framework for ECG denoising. The proposed framework consists of two networks: a generator that is composed of the optimized convolutional auto-encoder (CAE) and a discriminator that is composed of four convolution layers and one full connection layer. As the convolutional layers of CAE can preserve spatial locality and the neighborhood relations in the latent higher-level feature representations of ECG signal, and the skip connection facilitates the gradient propagation in the denoising training process, the trained denoising model has good performance and generalization ability. The extensive experimental results on MIT-BIH databases show that for single noise and mixed noises, the average signal-to-noise ratio (SNR) of denoised ECG signal is above 39 dB, and it is better than that of the state-of-the-art methods. Furthermore, the denoised classification results of four cardiac diseases show that the average accuracy increased above 32 $\%$ under multiple noises under SNR=0 dB. So, the proposed method can remove noise effectively as well as keep the details of the features of ECG signals. [ABSTRACT FROM AUTHOR]