Search Results (107)

Radio frequency fingerprint identification (RFFI) leverages the unique characteristics of radio signals resulting from inherent hardware imperfections for identification, making it essential for applications in telecommunications, cybersecurity, and surveillance. Despite the advancements brought by deep learning in enhancing RFFI accuracy, challenges persist in model deployment, particularly when transferring RFFI models across different receivers. Variations in receiver hardware can lead to significant performance declines due to shifts in data distribution. This paper introduces the source-free cross-receiver RFFI (SCRFFI) problem, which centers on adapting pre-trained RF fingerprinting models to new receivers without needing access to original training data from other devices, addressing concerns of data privacy and transmission limitations. We propose a novel approach called contrastive source-free cross-receiver network (CSCNet), which employs contrastive learning to facilitate model adaptation using only unlabeled data from the deployed receiver. By incorporating a three-pronged loss function strategy—minimizing information entropy loss, implementing pseudo-label self-supervised loss, and leveraging contrastive learning loss—CSCNet effectively captures the relationships between signal samples, enhancing recognition accuracy and robustness, thereby directly mitigating the impact of receiver variations and the absence of source data. Our theoretical analysis provides a solid foundation for the generalization performance of SCRFFI, which is corroborated by extensive experiments on real-world datasets, where under realistic noise and channel conditions, that CSCNet significantly improves recognition accuracy and robustness, achieving an average improvement of at least 13% over existing methods and, notably, a 47% increase in specific challenging cross-receiver adaptation tasks. Full article

Celosiae Argentea Semen (CAS), derived from Celosia argentea L., is traditionally used in Korean and Chinese medicine to treat eye disorders and liver heat and is recognized in official Pharmacopeias. In contrast, Celosiae Cristatae Semen (CCS), despite its frequent presence in the market, is not officially listed. The morphological and chemical similarities between the two pose challenges for accurate identification. This study presents an integrative method combining digital image analysis and high-performance thin-layer chromatography coupled with mass spectrometry (HPTLC-MS) to differentiate CAS from CCS. Digital microscopy and ImageJ analysis showed that CCS has a projection area over twice that of CAS. Chemically, an optimized HPTLC method using ethyl acetate, methanol, water, and formic acid revealed distinct fingerprint patterns under UV 366 nm and white light. Notably, celosin F was exclusively detected in CAS, while celosin H, J, and K were characteristic of CCS. ESI-TOF-MS analysis confirmed these markers, resolving an overlap in R_F values. Repeatability tests showed total SDs of sucrose for intra-day, inter-day, and inter-analysis precision were 0.006, 0.004, and 0.005, respectively, confirming method reliability. This combined approach offers a rapid, reliable, and practical tool for distinguishing these two medicinal seeds, supporting enhanced quality control and regulatory standardization in pharmaceutical applications. Full article

Radio Frequency Machine Learning (RFML) has in recent years become a popular method for performing a variety of classification tasks on received signals. Among these tasks is Specific Emitter Identification (SEI), which seeks to associate a received signal with the physical emitter that transmitted it. Many different model architectures, including individual classifiers and ensemble methods, have proven their capabilities for producing high accuracy classification results when performing SEI. Though the works studying different model architectures report on successes, there is a notable absence regarding the examination of systemic failures and negative traits associated with learned behaviors. This work studies those failure patterns for a 64-radio SEI classification problem by isolating common patterns in incorrect classification results across multiple model architectures and two distinct control variables: Signal-to-Noise Ratio (SNR) and the quantity of training data utilized. This work finds that many of the RFML-based models devolve to selecting from amongst a small subset of classes (≈10% of classes) as SNRs decrease and that observed errors are reasonably consistent across different SEI models and architectures. Moreover, our results validate the expectation that ensemble models are generally less brittle, particularly at a low SNR, yet they appear not to be the highest-performing option at a high SNR. Full article

Edge computing (EC) faces unique security threats due to its distributed architecture, resource-constrained devices, and diverse applications, making it vulnerable to data breaches, malware infiltration, and device compromise. The mitigation strategies against EC data security threats include encryption, secure authentication, regular updates, tamper-resistant hardware, and lightweight security protocols. Physical Unclonable Functions (PUFs) are digital fingerprints for device authentication that enhance interconnected devices’ security due to their cryptographic characteristics. PUFs produce output responses against challenge inputs based on the physical structure and intrinsic manufacturing variations of an integrated circuit (IC). These challenge-response pairs (CRPs) enable secure and reliable device authentication. Our work implements the Arbiter PUF (APUF) on Altera Cyclone IV FPGAs installed on the ALINX AX4010 board. The proposed APUF has achieved performance metrics of 49.28% uniqueness, 38.6% uniformity, and 89.19% reliability. The robustness of the proposed APUF against machine learning (ML)-based modeling attacks is tested using supervised Support Vector Machines (SVMs), logistic regression (LR), and an ensemble of gradient boosting (GB) models. These ML models were trained over more than 19K CRPs, achieving prediction accuracies of 61.1%, 63.5%, and 63%, respectively, thus cementing the resiliency of the device against modeling attacks. However, the proposed APUF exhibited its vulnerability to Multi-Layer Perceptron (MLP) and random forest (RF) modeling attacks, with 95.4% and 95.9% prediction accuracies, gaining successful authentication. APUFs are well-suited for device authentication due to their lightweight design and can produce a vast number of challenge-response pairs (CRPs), even in environments with limited resources. Our findings confirm that our approach effectively resists widely recognized attack methods to model PUFs. Full article

Device identification based on radio frequency fingerprinting is widely used to improve the security of Internet of Things systems. However, noise and acquisition inconsistencies in raw radio frequency signals can affect the effectiveness of classification, identification and authentication algorithms used to distinguish Bluetooth devices. This study investigates how the RF signal preprocessing techniques affect the performance of a support vector machine classifier based on radio frequency fingerprinting. Four options derived from an RF signal preprocessing technique are evaluated, each of which is applied to the raw radio frequency signals in an attempt to improve the consistency between signals emitted by the same Bluetooth device. Experiments conducted on raw Bluetooth signals from twentyfour smartphone radios from two public databases of RF signals show that selecting an appropriate RF signal preprocessing approach can significantly improve the effectiveness of a support vector machine classifier-based algorithm used to discriminate Bluetooth devices. Full article

Objective: Falciparum malaria is a major global health concern, affecting more than half of the world’s population and causing over half a million deaths annually. Red cell invasion is a crucial step in the parasite’s life cycle, where the parasite invade human erythrocytes to sustain infection and ensure survival. Two parasite proteins, Apical Membrane Antigen 1 (AMA-1) and Rhoptry Neck Protein 2 (RON2), are involved in tight junction formation, which is an essential step in parasite invasion of the red blood cell. Targeting the AMA-1 and RON2 interaction with inhibitors halts the formation of the tight junction, thereby preventing parasite invasion, which is detrimental to parasite survival. This study leverages machine learning (ML) to predict potential small molecule inhibitors of the AMA-1–RON2 interaction, providing putative antimalaria compounds for further chemotherapeutic exploration. Method: Data was retrieved from the PubChem database (AID 720542), comprising 364,447 inhibitors and non-inhibitors of the AMA-1–RON2 interaction. The data was processed by computing Morgan fingerprints and divided into training and testing with an 80:20 ratio, and the classes in the training data were balanced using the Synthetic Minority Oversampling Technique. Five ML models developed comprised Random Forest (RF), Gradient Boost Machines (GBMs), CatBoost (CB), AdaBoost (AB) and Support Vector Machine (SVM). The performances of the models were evaluated using accuracy, F1 score, and receiver operating characteristic—area under the curve (ROC-AUC) and validated using held-out data and a y-randomization test. An applicability domain analysis was carried out using the Tanimoto distance with a threshold set at 0.04 to ascertain the sample space where the models predict with confidence. Results: The GBMs model emerged as the best, achieving 89% accuracy and a ROC-AUC of 92%. CB and RF had accuracies of 88% and 87%, and ROC-AUC scores of 93% and 91%, respectively. Conclusions: Experimentally validated inhibitors of the AMA-1–RON2 interaction could serve as starting blocks for the next-generation antimalarial drugs. The models were deployed as a web-based application, known as PLASMOpred. Full article

Butyrylcholinesterase (BChE), plays a critical role in alleviating the symptoms of Alzheimer’s disease (AD) by regulating acetylcholine levels, emerging as an attractive target for AD treatment. This study employed a quantitative structure–activity relationship (QSAR) model based on ECFP4 molecular fingerprints with several machine learning algorithms (XGBoost, RF, SVM, KNN), among which the XGBoost model showed the best performance (AUC = 0.9740). A hybrid strategy integrating ligand- and structure-based virtual screening identified 12 hits from the Topscience core database, three of which were identified for the first time. Among them, piboserod and Rotigotine demonstrated the best BChE inhibitory potency (IC₅₀ = 15.33 μM and 12.76 μM, respectively) and exhibited favorable safety profiles as well as neuroprotective effects in vitro. Notably, Rotigotine, a marketed drug, was newly recognized for its anti-AD potential, with further enzyme kinetic analyses revealing that it acts as a mixed-type inhibitor in a non-competitive mode. Fluorescence spectroscopy, molecular docking, and molecular dynamics simulations further clarified their binding modes and stability. This study provides an innovative screening strategy for the discovery of BChE inhibitors, which not only identifies promising drug candidates for the treatment of AD but also demonstrates the potential of machine learning in drug discovery. Full article

To tackle the issue of unknown spoofing attacks in the Internet of Medical Things (IoMT), we put forward an iterative deep clustering model for blind RF fingerprint recognition. This model seamlessly combines a representation learning module and a clustering module, facilitating end—to—end training and optimization. Its parameters are updated according to an innovative loss function. Moreover, this model incorporates a noise—canceling self—encoder module to reduce noise and extract the noise—independent intrinsic fingerprints of devices. In comparison with other algorithms, the proposed model remarkably improves the blind recognition performance for the identification of unknown devices in the IoMT. Full article

The classification of Hepatitis C virus (HCV) NS3 inhibitors is essential for identifying potential antiviral agents through computational methods. This study aims to develop an optimized machine learning (ML) model using random forest (RF) and molecular fingerprints to accurately classify HCV NS3 inhibitors. A dataset of 965 molecules was retrieved from the ChEMBL database, and 290 bioactive compounds were selected for model training. Twelve molecular fingerprint descriptors were tested, and the CDK graph-only fingerprint yielded the best performance. In addition to RF, performance comparisons of other classifiers such as instance-based k-nearest neighbor (IBk), logistic regression (LR), AdaBoost, and OneR were conducted using WEKA with various molecular fingerprint descriptors. The optimized RF model achieved an accuracy of 89.6552%, a mean absolute error (MAE) of 0.2114, a root mean square error (RMSE) of 0.3304, and a Matthews correlation coefficient (MCC) of 0.7950 on the test set. These results highlight the effectiveness of optimized molecular fingerprints in enhancing virtual screening (VS) for HCV inhibitors. This approach offers a data-driven method for drug discovery. Full article

Codonopsis Radix (CR), an important species of “medicine-food homology”, exhibits broad market prospects, underscoring the urgency and importance of research on its quality. This study specifically measured the alcohol-soluble extract and polysaccharide extract of 77 samples from mainstream producing areas of CR, which serve as key fractions for assessing its quality. Additionally, to gain a comprehensive understanding of the sensory characteristics of samples, the study employed electronic tongue technology to obtain sweetness values, used a colorimeter to determine yellowness values, and captured odor fingerprint information through an electronic nose (E-nose). In the data analysis phase, the study compared the accuracy of various regression prediction models, including Multiple Linear Regression (MLR), Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN). After comprehensive evaluation, an SVM algorithm was selected due to its superior prediction performance. To further enhance prediction accuracy, the study utilized a Particle Swarm Optimization (PSO) algorithm to optimize the SVM, resulting in a significant improvement in the prediction accuracy of sweetness values. In conclusion, regression prediction models for chemical composition and sensory information of CR based on an E-nose were established. It represents an enhancement of traditional morphological identification methods for Chinese medicinal herbs and provides new ideas and means for quality evaluation of CR. Furthermore, it offers a reference for quality evaluation of other similar Chinese medicinal herbs. Full article

The great interest in the rapid and reliable differentiation of extra virgin olive oil from other olive oil categories is directly related to its unique sensory characteristics and high market prices. The aim of the present study was to investigate the potential of FTIR as a rapid and non-invasive technique to discriminate extra virgin olive oil (EVOO) from other olive oil categories (virgin olive oil, ordinary, and lampante) based on the acquired spectral profile of olive oil. Spectral data were collected, pre-processed, and correlated by Random Forest (RF) analysis with the sensory category (EVOO vs. other) of olive oil samples, as defined by sensory analysis undertaken previously by trained panelists. The results showed that the application of Savitzky–Golay (S-G) smoothing with a second derivative (d = 2), second- and third-order polynomial (p = 2, p = 3), and window size (w) of 12 and 13 points achieved the highest accuracy (0.91) between the two classes of samples. Characteristic spectral bands of triacylglycerols related to the carbonyl groups present in triacylglycerols (C=O) located near 1744 cm⁻¹ (specific features: 1739, 1748, and 1751 cm⁻¹), the fingerprinting area 1250–1000 cm⁻¹ (specific features: 1088, 1094, 1116, 1123, 1124, 1158, 1162, 1236, 1240, and 1247 cm⁻¹), which correspond to CH bending, and 1680 cm⁻¹, which is associated with unsaturated aldehydes were observed to constitute the main basis of the discrimination of EVOO from the “other” class. The ability of the model to achieve high classification accuracy demonstrates the robustness of the FTIR spectral data combined with advanced machine learning techniques. Due to the lower cost and more rapid analysis time afforded by FTIR, this method provides promising perspectives for industrial olive oil classification. Full article

This research article proposes a new method for an enhanced Flexible Manufacturing System (FMS) using a combination of smart methods. These methods use a set of three technologies of Industry 4.0, namely Artificial Intelligence (AI), Digital Twin (DT), and Wi-Fi-based indoor localization. The combination tackles the problem of asset tracking through Wi-Fi localization using machine-learning algorithms. The methodology utilizes the extensive “UJIIndoorLoc” dataset which consists of data from multiple floors and over 520 Wi-Fi access points. To achieve ultimate efficiency, the current study experimented with a range of machine-learning algorithms. The algorithms include Support Vector Machines (SVM), Random Forests (RF), Decision Trees, K-Nearest Neighbors (KNN) and Convolutional Neural Networks (CNN). To further optimize, we also used three optimizers: ADAM, SDG, and RMSPROP. Among the lot, the KNN model showed superior performance in localization accuracy. It achieved a mean coordinate error (MCE) between 1.2 and 2.8 m and a 100% building rate. Furthermore, the CNN combined with the ADAM optimizer produced the best results, with a mean squared error of 0.83. The framework also utilized a deep reinforcement learning algorithm. This enables an Automated Guided Vehicle (AGV) to successfully navigate and avoid both static and mobile obstacles in a controlled laboratory setting. A cost-efficient, adaptive, and resilient solution for real-time tracking of assets is achieved through the proposed framework. The combination of Wi-Fi fingerprinting, deep learning for localization, and Digital Twin technology allows for remote monitoring, management, and optimization of manufacturing operations. Full article

The NLRP3 inflammasome, plays a critical role in the pathogenesis of rheumatoid arthritis (RA) by activating inflammatory cytokines such as IL1β and IL18. Targeting NLRP3 has emerged as a promising therapeutic strategy for RA. In this study, a multidisciplinary approach combining machine learning, quantitative structure–activity relationship (QSAR) modeling, structure–activity landscape index (SALI), docking, molecular dynamics (MD), and molecular mechanics Poisson–Boltzmann surface area MM/PBSA assays was employed to identify novel NLRP3 inhibitors. The ChEMBL database was used to retrieve compounds with known IC₅₀ values to train machine learning (ML) models using the Lazy Predict package. After data pre-processing, 401 non-redundant structures were selected for exploratory data analysis (EDA). PubChem and MACCS fingerprints were used to predict the inhibitory activities of the compounds. SALI was used to identify structurally similar compounds with significantly different biological activities. The compounds were docked using MOE to assess their binding affinities and interactions with key residues in NLRP3. The models were evaluated, and a comparative analysis revealed that the ensemble Random Forest (RF) model (PubChem fingerprints) with RMSE (0.731), R² (0.622), and MAPE (8.988) and bootstrap aggregating model (MACCS fingerprints) with RMSE (0.687), R² (0.666), and MAPE (9.216) on the testing set performed well, in accordance with the Organization for Economic Cooperation and Development (OECD) guidelines. Out of all docked compounds, the two most promising compounds (ChEMBL5289544 and ChEMBL5219789) with binding scores of −7.5 and −8.2 kcal/mol were further investigated by MD to evaluate their stability and dynamic behavior within the binding site. MD simulations (200 ns) revealed strong structural stability, flexibility, and interactions in the selected complexes. MM/PBSA binding free energy calculations revealed that van der Waals and electrostatic forces were the key drivers of the binding of the protein with ligands. The outcomes obtained can be used to design more potent and selective NLRP3 inhibitors as therapeutic agents for the treatment of inflammatory diseases such as RA. However, concerns related to the lack of large datasets, experimental validation, and high computational costs remain. Full article

Radio frequency (RF) fingerprinting is regarded as a promising solution to improve wireless security, especially in applications where resource-limited devices are employed. Unlike steady-state signals, such as preambles or data, the use of short-duration transient signals for RF fingerprinting offers distinct advantages for systems with low latency and low complexity requirements. One of the challenges associated with transient-based methods in practice is achieving high performance while utilizing low-cost receivers. In this study, we demonstrate for the first time that the performance of transient-based RF fingerprinting can be enhanced by designing the filter chain in a software defined radio (SDR) receiver, taking into account the relevant signal characteristics. The performance analysis is conducted using transient-based identification signals captured by the SDR receiver, focusing on the sampling rate and duration of the identification signal. In the experiments, signals collected from twenty IEEE 802.11 transmitters are used. Experimental results indicate that so long as the receiver filter parameters and the duration of the identification signal are properly determined, a high classification performance exceeding 92% can be achieved for transient-based RF fingerprinting, even at sampling rates approaching the Nyquist limit. Full article

Recent advancements in photonic integrated circuits (PICs) have paved the way for a new era of frequency-agile coherent radar systems. Unlike traditional all-electronic RF radar techniques, fully photonic systems offer superior performance, overcoming bandwidth limitations and noise degradation when operating across S (2–4 GHz), X (8–12 GHz), and K-band (12–40 GHz) frequencies. They also exhibit excellent phase noise performance, even at frequencies exceeding 20 GHz. However, current state-of-the-art PICs still suffer from high processing losses in the optical domain, necessitating careful design of the electrical RF domain. This study delves into the critical challenges of designing RF front-ends for microwave photonic radars, including stability, noise minimization, and intermodulation distortion reduction. To demonstrate the feasibility of the proposed design, a functional prototype is constructed, achieving a total power gain of 107 dB (radar system at 10 GHz) while minimizing signal noise degradation. Furthermore, a comprehensive demonstration of the RF front-end, encompassing both optical RF signal generation and experimental measurements of a rotor blade’s Doppler fingerprint with 0.5 Hz resolution, validates the proposed system’s performance. Full article