IET Signal Processing最新文献

Time–frequency–spatial (TFS) features play a crucial role in motor imagery electroencephalogram (EEG) classification. However, effectively leveraging these multidimensional features to enhance classification accuracy remains a significant challenge. Although feature selection techniques are widely used to extract informative TFS representations, most existing methods treat each dimension independently, overlooking the intrinsic grouping and hierarchical relationships among them. To address this limitation, this paper proposes a hierarchical feature recalibration network (HFRNet) that explicitly models intergroup dependencies and the hierarchical structure of TFS features, thereby substantially improving motor imagery EEG classification performance. HFRNet employs a two-layer weighting mechanism for hierarchical feature recalibration, followed by a classification module. In the first weighting layer, spatial features within each time–frequency segment are grouped and represented as feature maps. Channel-wise dependencies are captured through squeeze-and-excitation operations to learn channel weights, which are then used to rescale each feature map. In the second weighting layer, the recalibrated features are reorganized across time windows and further refined through a similar recalibration process. Finally, in the classification block, the refined features are flattened, concatenated into a single feature vector, and passed through dropout and fully connected (FC) layers for classification. Extensive experiments conducted on five motor imagery datasets demonstrate that the proposed HFRNet achieves the best overall performance, with an average accuracy (F1 score) of 81.03% (0.7931). Comparative evaluations against 30 feature selection methods and recent state-of-the-art approaches further confirm the superior effectiveness and robustness of the proposed model.

In this paper, a nonorthogonal multiple access (NOMA) network is considered to assist ground users’ communications with two reconfigurable intelligent surfaces (RISs) and one unmanned aerial vehicle (UAV) technologies. In this scheme, UAV acts as a relay while the first RIS aids ground-to-air (G2A) communication and the second RIS helps the air-to-ground (A2G) communication. Under the outage probability (OP) constraint, the problem is formulated to maximize the data rate of the ground users by optimizing the transmission powers of base station (BS) and UAV, power allocation coefficients to the users and phase shift of the reflecting elements (REs) of RISs. An iterative algorithm is proposed using artificial bee colony (ABC) method for solving the problem and improving the network performance. Proposed algorithm is validated by Monte Carlo simulations, and the impact of system parameters is investigated on the network performance. Numerical results clarify that by two RISs utilization, the performances of the network are approximately improved 25% and 40% in terms of the network data rate and total OP, respectively, in comparison with the bench mark algorithms.

Depressions affects the entire nervous system and, in turn, human behavior. Electroencephalogram (EEG) signal classification of depression datasets using traditional methods takes time. Non-invasive EEG signals provide valuable insights into this mental health condition’s neural patterns and abnormalities. The proposed ENS model classifies better than many other machine learning models on these EEG datasets. Thus, it will be used to investigate the dataset and classify it as normal or depressed. The ENS model reduces dimensions after extracting features, and multiple classifiers classify the dataset. The proposed work attains a maximum classification accuracy of 97%. In order to validate the hardware’s computational efficiency, the proposed method was implemented in FPGA, and performance analyses were performed on various multiply-accumulate (MAC) units. Overall performance of the proposed work is improved to 98.8% compared to the conventional approach.

With the rapidly increasing global climate change, rainfall forecasting is highly valuable for water resource planning, irrigation management, flood control, etc. Capturing the nonlinear rainfall behavioris very complex. One of the greatest difficulties in predicting rainfall is the treatment of extremely nonlinear and noisy rainfall patterns. To address this issue, the current research utilized a hybrid data preprocessing technique that integrates a signal processing technique called empirical model decomposition (EMD) with the Hilbert transform (HT) for addressing noise as well as highly nonlinear rainfall time series and the famous machine learning (ML) algorithm, namely the decision tree (DT) and artificial neural network (ANN). For comparative purposes, the hybrid EMD-based model and the lag-time-based traditional model have been developed. For model construction, the data are split into two phases: training (80%) and testing (20%). The model was validated for different performance metrics, such as mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and R-squared (R²). The proposed model improved R² by 52.14% and 21.54% over the two hybrid approaches and by 130.72% and 203.77% over the two traditional lag-time-based approaches, respectively. The outcomes reveal that prediction accuracy improves with the use of EMD with the HT compared to EMD by itself and the conventional lag-time methodologies employed here. The realization encourages researchers to implement this technique in other geological regions for rainfall forecasting.

Today, modern radar systems increase their target detection capabilities by processing pulses coherently. If these radars do not take precautions against the Doppler frequency shift caused by the moving target, their performance will decrease. In addition, if these radars do not change their parameters from pulse to pulse, modern jammers will cause them to produce false alarms. In this article, we add Doppler compensation capability to the smart binary phase-coding (SBPC) method. The proposed SBPC-Doppler method is recommended to facilitate radar detection of moving targets and suppress repetitive range deception techniques. In the simulations, the traditional approach in which the same code is used without changing from pulse to pulse, and the approaches using code sets obtained by the old SBPC and the new SBPC-Doppler methods in intrapulse modulation are compared. The results show that the proposed SBPC-Doppler method can significantly improve the isolation against deception jamming and the moving target detection capability simultaneously.

Multipoint optimal minimum entropy deconvolution adjusted (MOMEDA) leverages its unique multiobjective optimization capability to effectively enhance specific fault-related features in signals. However, its core performance relies on the accurate prior definition of the fault period T. This inherent dependence on precise prior configuration limits its practical application. Moreover, the fitness function of the existing optimization algorithm is only designed for a specific single rotating component in the process of MOMEDA’s parameter selection. Therefore, an improved MOMEDA method based on a new fitness function is proposed. This approach begins with the design of a comprehensive fitness function, CHK, that integrates both impulsive and periodic characteristics. A dynamic weighting mechanism adaptively balances the fault features of diverse targets, enabling effective identification of various fault patterns, including those in bearings and gears. Furthermore, a particle swarm optimization (PSO) algorithm is employed to adaptively optimize the deconvolution period T, a critical parameter in MOMEDA. This optimization algorithm employs the proposed CHK indicator as the fitness function. The effectiveness and generalization of the proposed method are validated through experiments of bearing and gear fault diagnosis. Finally, the experimental results demonstrate that the proposed method is able to accurately extract subtle fault features from different rotating machinery. It also shows that this method exhibits significant advantages regarding anti-interference capabilities and application scope, when compared with SK-based MOMEDA, envelope entropy-based MOMEDA, and peak factor of envelope spectrum-based MOMEDA.

Eigenvalue detection is extensively utilized in numerous applications, including spectrum sensing in cognitive radio. However, the characteristics of the Tracy–Widom ($$) distribution and its lack of accurate close-form expression limit the detection performance based on the extreme eigenvalue. The paper introduces an eigenvector correlation (EVC) based detection. It extracts the eigenvalue of a weak component from the Marchenko–Pastur law (MP-law) bulk by combining a priori known eigenvector-correlated signal. The extracted eigenvalue then follows an easily analyzable and tractable Gaussian distribution according to the random matrix theory. The proposed method significantly improves sensing performance, which is theoretically analyzed and compared with traditional maximum eigenvalue-based detection (MED) and trace-based sensing. In addition, a cumulative EVC (CUM–EVC) is further developed for the multiple eigen-component signal. Numerical results are presented to validate the theoretical analysis and demonstrate the reliability of the proposed detectors.

Noncooperative wireless networks are characterized by decentralized control, the absence of collaboration among nodes, and unpredictable environmental factors, all of which present significant challenges for network topology inference due to the limited availability of network information. To address this issue, we propose a novel topology inference methodology that leverages the time series of packet transmission times without relying on packet decoding. Our approach begins by converting signal transmission times into discrete time series data, which serves as the foundation for topology inference, transforming the connectivity problem into a causality problem. Rather than using traditional linear Granger causality (GC), we propose a novel architecture called difference neural GC (DNGC), which excels in learning network topology from sampled time series data without requiring access to the underlying protocol. By utilizing a hierarchical penalty and a differencing approach as adaptive weights, DNGC effectively captures the dynamic and nonlinear relationships between neighboring time steps in the collected sequence. Extensive simulations demonstrate that DNGC outperforms existing GC-based methods, particularly when observation time is limited.

Compared to human speech and multimedia audio, the amplitude of lung respiratory sounds is extremely weak, with minimal differences between the respiratory sounds of various lung diseases. Traditional methods, such as Mel-frequency cepstral coefficients (MFCCs) and the Fourier transform, struggle to accurately extract respiratory sound characteristics of different lung diseases. Therefore, this article innovatively employs an improved C₀ complexity method to analyze the subtle differences in respiratory sounds, identifying the optimal tuning factor value that captures more local information and better characterizes the nonlinear stochastic properties of respiratory sounds. Subsequently, the second component of the MFCC was extracted and fused with the improved C₀ complexity and short-term energy to propose a novel Mel-energy complexity (MEC) feature. Finally, the fuzzy C-means (FCM) clustering method was employed to process the MEC features, detecting the periodic endpoints of respiratory sounds, specifically, the endpoints of respiratory sounds during inhalation and exhalation, thereby laying a technical foundation for further research on respiratory sound recognition and visual management. To evaluate performance, we established a clinically collected respiratory sound database containing respiratory sounds from 154 individuals across four categories, including asthma and chronic obstructive pulmonary disease (COPD), with sounds recorded from eight chest locations per participant. The proposed method achieved average accuracy and F1 scores of 71% and 72%, respectively, outperforming multiple respiratory sound endpoint detection (RSED) approaches while demonstrating high robustness, particularly in processing pulmonary crackles.

Surface electromyography (sEMG) has been used for decades to diagnose movement and neuromuscular disorders; however, sEMG signals are noisy and interfered with, and the nonstationary, nonlinear nature of sEMG signals complicates their use for diagnostic purposes. However, existing denoising methods often sacrifice the original signal, thereby losing practical physiological details, which makes them clinically less applicable. To overcome these challenges, we introduce generalized successive variational mode decomposition (GSVMD), an advanced denoising technique that preserves signal integrity over all frequencies. GSVMD decouples Successive variational mode decomposition (SVMD) for adaptive signal decomposition, Soft interval thresholding (SIT) for focused noise reduction, and attention mechanisms to focus on clinically relevant signal components. GVSMD was evaluated using 4-channel sEMG data from 12 healthy participants and 24 stroke patients, demonstrating superior performance compared to traditional methods, with a significant increase in SNR and R² values. Its robustness was confirmed by statistical validation (p < 0.001, p < 0.05). Taken together, these findings highlight GSVMD’s potential for real-time clinical diagnostics and its potential application in a wide range of patient groups and conditions.