Electronics Letters最新文献

Semantic labelling of remote sensing images is crucial for various remote sensing applications. However, the dense distribution of man-made and natural objects with similar colours and geographic proximity poses challenges for achieving consistent and accurate labelling results. To address this issue, a novel deep learning model incorporating an octave convolutional neural networks (CNNs) within an end-to-end U-shaped architecture is presented. The approach differs from conventional CNNs in that it employs octave convolutions instead of standard convolutions. This strategy serves to minimize low-frequency information redundancy while maintaining segmentation accuracy. Furthermore, coordination attention is introduced in the encoder module to enhance the network's ability to extract useful features, focusing on spatial and channel dependencies within the feature maps. This attention mechanism enables the network to better capture channel, direction, and location information. In conclusion, the U-shaped network is engineered with a completely symmetric structure that employs skip connections to merge low-resolution information, used for object class recognition, with high-resolution information to enable precise localization. This configuration ultimately improves segmentation accuracy. Experimental results on two public datasets demonstrate that our U-ONet achieves state-of-the-art performance, making it a compelling choice for remote sensing image semantic labelling applications.

An anomalous transfer of modulation from a modulated microwave beam ($�F_2$) to an unmodulated one ($�F_1$) has been experimentally demonstrated through accurate delay time measurements. These results are analyzed according to two possible interpretations: one based on an electromagnetic approach as reported elsewhere, and the other based on a purely stochastic model. This latter model, made up of random zig-zag paths, is the focus of the current paper. It is demonstrated that it is possible to obtain a plausible description of the experimental data.

The problem of user equipment positioning based on radio signals is considered via deep learning. As in most supervised-learning tasks, a critical aspect is the availability of a relevant dataset to train a model. However, in a cellular network, the data-collection step may induce a high communication overhead. As a result, to reduce the required size of the dataset, it may be interesting to carefully choose the positions to be labelled and to be used in the training. Therefore, an active learning approach for efficient data collection is proposed. It is first shown that significant gains (both in terms of positioning accuracy and size of the required dataset) can be obtained for the considered positioning problem using a genie. This validates the interest of active learning for positioning. Then, a practical method is proposed to approximate this genie.

In this letter, an improved gated linear unit (GLU) structure for end-to-end (E2E) speech enhancement is proposed. In the U-Net structure, which is widely used as the foundational architecture for E2E deep neural network-based speech denoising, the input noisy speech signal undergoes multiple layers of encoding and is compressed into essential potential representative information at the bottleneck. The latent information is then transmitted to the decoder stage for the restoration of the target clean speech. Among these approaches, CleanUNet, a prominent state-of-the-art (SOTA) method, enhances temporal attention in latent space by employing multi-head self-attention. However, unlike the approach of applying the attention mechanism to the potentially compressed representative information of the bottleneck layer, the proposed method instead assigns the attention module to the GLU of each encoder/decoder block layer. The proposed method is validated by measuring short-term objective speech intelligibility and sound quality. The objective evaluation results indicated that the proposed method using residual-attention GLU outperformed existing methods using SOTA models such as FAIR-denoiser and CleanUNet across signal-to-noise ratios ranging from 0 to 15 dB.

As 5G and early-stage network technologies mature, a wealth of data and experience is accumulated. The ubiquitous sensing and communication network (USCN) can leverage these existing communication infrastructures to maximize resource utilization and reduce the costs of redeployment through integration with modern 6G and beyond 6G networks. However, implementing future integrated sensing and communication (ISAC) functions in USCN may possess different hardware capabilities and communication protocols, presenting significant challenges. This is due to the lack of a unified framework for managing and optimizing these heterogeneous resources, as well as the absence of reasonable performance metrics. This letter proposes a dynamic pricing coordinated resource allocation (DPCRA) mechanism to realize the emerging ISAC in legacy USCNs. The DPCRA is designed to optimize the spectrum efficiency (SE) of USCN and manages resource interactions between heterogeneous devices through price variables, thereby resolving resource conflicts and deadlocks. The novelties include: (1) employing a price mechanism to foster resource cooperation in non-cooperative networks; (2) introducing an SE metric incorporating an information similarity factor for optimization; (3) dynamically selecting resources to enhance SE. Simulations demonstrate that the DPCRA effectively resolves device deadlock while maintaining high SE levels, promoting device collaboration.

To address the issue of radar detection performance degradation caused by interference in complex electromagnetic environments, this paper proposes a generative waveform design method and demonstrates its advantages. This method fully exploits waveform degrees of freedom to achieve a larger design space, generating a multitude of agile waveform parameters within the constraint envelope through algorithmic processes. This approach enables the radar to maintain detection performance in complex electromagnetic environments. Compared with the traditional waveform design method, the simulation experiment verifies the effectiveness of the generative waveform design method.

The detection of defects on the copper chain in the production process of tobacco cutters is crucial for ensuring product quality. Traditional defect detection methods often rely on spatial domain image analysis, which not only has a large computational load but also performs poorly in handling high-frequency noise and complex backgrounds. To address this issue, this paper proposes a novel neural network model based on frequency domain analysis, called frequency domain attention network. This network first utilizes discrete cosine transform to transform the image from the spatial domain to the frequency domain, effectively reducing computational complexity and improving processing speed. Subsequently, through the innovative frequency domain attention module, the network automatically identifies and enhances key discriminative features in the frequency domain, thereby strengthening the model's ability to identify defects. Finally, the frequency domain attention map, after feature extraction and integration, is inputted into the coupling detection head to achieve high-precision defect detection. The experimental results show that our method outperforms the SOTA method with an increase of 0.03 in AP and 21 in FPS.

A twice-input variable-resolution single-side switching scheme without reset energy is proposed for successive approximation register (SAR) analogue-to-digital converters (ADCs). The proposed switching scheme is based on semi-resting DAC technology to design a four-array architecture capable of handling twice the swing of the signal input while reducing the capacitor array size. The technique of top-plate sampling and monotonic shift is utilized so that no switching energy is generated for the first three comparisons. The proposed scheme utilizes full-capacitor split, bridged switch, and C-2C dummy capacitor technology, which reduces average switching energy consumption by 99.8% compared to conventional schemes and enables ADC variable resolution function, making the SAR ADC more suitable for IoT applications.

The data scale keeps growing by leaps and the majority of it is high-dimensional imbalanced data, which is hard to classify. Data missing often happens in software which further aggravates the difficulty of classifying the data. In order to resolve high-dimensional imbalanced mixed-variables missing data classification problem, a novel method based on particle swarm optimization is developed. It has three original components including multiple feature selection, mixed attribute imputation, and quantum oversampling. Multiple feature selection uses a two-stage strategy to obtain stable relevant features. Mixed attribute imputation separates features into continuous and discrete features and fills missing values with different models. Quantum oversampling chooses instances to balance data based on the quantum operator. Furthermore, particle swarm optimization is employed to optimize the parameters of the components to obtain preferable classification results. Six representative classification datasets, six typical algorithms, and four measures are taken to conduct exhaust experiments, and results indicate that the proposed method is superior to the comparison algorithms.

The static frequency converter (SFC) in a pumped storage power plant often causes harmonic problems in the dragging processes, which may lead to the false operation of automatic devices in the power station, and even damage to the power equipment. These harmonics caused by SFC contain both characteristic and non-characteristic degrees, and the components are more complex. Hence, the existing harmonic suppression methods using active power filter or passive power filter compensation all have some problems. The SFC starting control strategy based on linear active disturbance rejection control (LADRC) is proposed here to reduce the non-characteristic harmonics. First, the factors affecting the non-characteristic harmonic content are analysed, and a mathematical model of the conventional SFC starting transfer function is established. On this basis, the LADRC controller is used as the speed loop of SFC starting to replace the proportional integral (PI) controller. The stability is judged by drawing the Bode plot and the zero-pole plot. Finally, a Matlab/Simulink simulation model of LADRC-based SFC starting is established in a pumped storage power plant with actual parameters, and simulation accurate measurements verify the effectiveness of the proposed control strategy for non-characteristic harmonic current suppression.