Low-sample Data Research Articles

AI-enhanced backpack with double frequency-up conversion vibration energy converter for motion recognition and extended battery life

With growing interest in physical activities, wearable device sales have been increasing each year due to their help in tracking movement. However, keeping these devices running for a long time is still a key challenge. This paper presents an AI-enhanced backpack equipped with a hybrid vibration energy converter that integrates piezoelectric (PE) and electromagnetic (EM) technologies, featuring a double frequency-up conversion (FUC) mechanism. The PE unit enables efficient motion recognition, while the EM unit extends battery life via kinetic energy harvesting. The double FUC mechanism boosts human motion signal strength and resolution by converting sub-Hertz frequencies to hundreds of Hertz. This makes the device particularly effective in environments with performance constraints and interference. By adopting deep learning techniques, the system maintains a high motion recognition accuracy of 97.33 % even under low data sampling rates and significant noise interference. The EM unit effectively harvests kinetic energy, generating an average output power of 4.5 mW during activities such as running at 4.5 Hz, despite of the device’s compact size. Additionally, a power management circuit with human motion-stimulated switch optimizes battery life by managing the system’s sleep and active states. This innovative energy management approach ensures that during human activity, the system conserves energy by remaining in sleep state for at least 30 % of the time, significantly extending battery life and enhancing the backpack’s suitability for outdoor applications.

Modeling the Impacts of Climate Change on Potential Distribution of Betula luminifera H. Winkler in China Using MaxEnt

Betula luminifera H. Winkler, a fast-growing broad-leaved tree species native to China’s subtropical regions, possesses significant ecological and economic value. The species’ adaptability and ornamental characteristics make it a crucial component of forest ecosystems. However, the impacts of global climate change on its geographical distribution are not well understood, necessitating research to predict its potential distribution shifts under future climate scenarios. Our aims were to forecast the impact of climate change on the potential suitable distribution of B. luminifera across China using the MaxEnt model, which is recognized for its high predictive accuracy and low sample data requirement. Geographical coordinate data of B. luminifera distribution points were collected from various databases and verified for redundancy. Nineteen bioclimatic variables were selected and screened for correlation to avoid overfitting in the model. The MaxEnt model was optimized using the ENMeval package, and the model accuracy was evaluated using the Akaike Information Criterion Correction (delta.AICc), Training Omission Rate (OR10), and Area Under the Curve (AUC). The potential distribution of B. luminifera was predicted under current and future climate scenarios based on the Shared Socio-economic Pathways (SSPs). The optimized MaxEnt model demonstrated high predictive accuracy with an AUC value of 0.9. The dominant environmental variables influencing the distribution of B. luminifera were annual precipitation, minimum temperature of the coldest month, and standard deviation of temperature seasonality. The potential suitable habitat area and its geographical location were predicted to change significantly under different future climate scenarios, with complex dynamics of habitat expansion and contraction. The distribution centroid of B. luminifera was also predicted to migrate, indicating a response to changing climatic conditions. Our findings underscore the importance of model optimization in enhancing predictive accuracy and provide valuable insights for the development of conservation strategies and forest management plans to address the challenges posed by climate change.

Variable horizon-based predictive energy management strategy for plug-in hybrid electric vehicles and determination of a suitable predictive horizon

For plug-in hybrid electric vehicles, the predictive energy management can be implemented in real-time with a short-term prediction of future driving condition. However, the completely precise prediction of future driving conditions and the determination of a reasonable predictive horizon remain quite difficult. To solve the above problems, we propose a variable horizon-based predictive energy management control method with adaptive SOC constraint, and more importantly, a reasonable predictive horizon is determined to balance global energy optimality and the cost of data sampling for the connected vehicles. Firstly, an integrated velocity predictor is proposed by combining Markov model and speed limits. Considering that the effect of information amount on prediction accuracy is opposite to that of predictive horizon, a new evaluation metric, namely predictive information entropy, is proposed to objectively evaluate the velocity prediction accuracy under different amounts of available information. Then, a variable horizon-based predictive energy management strategy is developed with the dynamically updated reference SOC. By comprehensively considering the effects of predictive horizon on prediction accuracy, real-time application and global optimality, a suitable predictive horizon should be set to 20s–30s to achieve a better fuel economy with a relatively low data sampling cost.

CoSev: Data-Driven Optimizations for COVID-19 Severity Assessment in Low-Sample Regimes

Given the pronounced impact COVID-19 continues to have on society—infecting 700 million reported individuals and causing 6.96 million deaths—many deep learning works have recently focused on the virus’s diagnosis. However, assessing severity has remained an open and challenging problem due to a lack of large datasets, the large dimensionality of images for which to find weights, and the compute limitations of modern graphics processing units (GPUs). In this paper, a new, iterative application of transfer learning is demonstrated on the understudied field of 3D CT scans for COVID-19 severity analysis. This methodology allows for enhanced performance on the MosMed Dataset, which is a small and challenging dataset containing 1130 images of patients for five levels of COVID-19 severity (Zero, Mild, Moderate, Severe, and Critical). Specifically, given the large dimensionality of the input images, we create several custom shallow convolutional neural network (CNN) architectures and iteratively refine and optimize them, paying attention to learning rates, layer types, normalization types, filter sizes, dropout values, and more. After a preliminary architecture design, the models are systematically trained on a simplified version of the dataset-building models for two-class, then three-class, then four-class, and finally five-class classification. The simplified problem structure allows the model to start learning preliminary features, which can then be further modified for more difficult classification tasks. Our final model CoSev boosts classification accuracies from below 60% at first to 81.57% with the optimizations, reaching similar performance to the state-of-the-art on the dataset, with much simpler setup procedures. In addition to COVID-19 severity diagnosis, the explored methodology can be applied to general image-based disease detection. Overall, this work highlights innovative methodologies that advance current computer vision practices for high-dimension, low-sample data as well as the practicality of data-driven machine learning and the importance of feature design for training, which can then be implemented for improvements in clinical practices.

Effective Voting-Based Ensemble Learning for Segregated Load Forecasting With Low Sampling Data

Effective Voting-Based Ensemble Learning for Segregated Load Forecasting With Low Sampling Data

Metagenomic Methods for Addressing NASA's Planetary Protection Policy Requirements on Future Missions: A Workshop Report.

Molecular biology methods and technologies have advanced substantially over the past decade. These new molecular methods should be incorporated among the standard tools of planetary protection (PP) and could be validated for incorporation by 2026. To address the feasibility of applying modern molecular techniques to such an application, NASA conducted a technology workshop with private industry partners, academics, and government agency stakeholders, along with NASA staff and contractors. The technical discussions and presentations of the Multi-Mission Metagenomics Technology Development Workshop focused on modernizing and supplementing the current PP assays. The goals of the workshop were to assess the state of metagenomics and other advanced molecular techniques in the context of providing a validated framework to supplement the bacterial endospore-based NASA Standard Assay and to identify knowledge and technology gaps. In particular, workshop participants were tasked with discussing metagenomics as a stand-alone technology to provide rapid and comprehensive analysis of total nucleic acids and viable microorganisms on spacecraft surfaces, thereby allowing for the development of tailored and cost-effective microbial reduction plans for each hardware item on a spacecraft. Workshop participants recommended metagenomics approaches as the only data source that can adequately feed into quantitative microbial risk assessment models for evaluating the risk of forward (exploring extraterrestrial planet) and back (Earth harmful biological) contamination. Participants were unanimous that a metagenomics workflow, in tandem with rapid targeted quantitative (digital) PCR, represents a revolutionary advance over existing methods for the assessment of microbial bioburden on spacecraft surfaces. The workshop highlighted low biomass sampling, reagent contamination, and inconsistent bioinformatics data analysis as key areas for technology development. Finally, it was concluded that implementing metagenomics as an additional workflow for addressing concerns of NASA's robotic mission will represent a dramatic improvement in technology advancement for PP and will benefit future missions where mission success is affected by backward and forward contamination.

LNA linearization solution for direct conversion receiver using under-sampling technique in reference receiver

The paper proposes a method for linearizing low noise amplifiers (LNAs) in multichannel direct conversion receivers. The proposed direct conversion receiver (DCR) uses a linear reference receiver to extract distortion information, which is then fed to an adaptive circuit for linearizing the main channel signal. The proposed DCR differs from prior LNA linearization techniques in that the reference channel in the proposed DCR uses analog to digital converter (ADC) with an undersampling technique to extract reference information. The low-speed ADC also serves as a downconverter, shifting radio frequency (RF) signal to baseband and allowing for all further linearization processing to be performed digitally at a low-sampling data rate. This significantly reduces cost, design complexity, and energy consumption. The effectiveness of the proposed design is theoretically verified through MATLAB simulation and practically measured for a 65 Mhz band of ultra-high frequency (UHF) DCR capable of simultaneously receiving four 16–QAM channels of the same bandwidth of 4 Mhz. The MATLAB software simulation results show that the proposed approach significantly improved the signal-to-noise and distortion ratio (SNDR) for the channel of interest by approximately 30 dB in the worst distorted channel. For hardware implementation, the distorted signals are sampled from a commercial LNA (ZFL–500LN+) by a customized FPGA board. Results from measurements show an improvement of 14.6% for error vector magnitude (EVM) in a strong distortion scenario of 16–QAM modulation signal.

Research on Data-Driven Optimal Scheduling of Power System

The uncertainty of output makes it difficult to effectively solve the economic security dispatching problem of the power grid when a high proportion of renewable energy generating units are integrated into the power grid. Based on the proximal policy optimization (PPO) algorithm, a safe and economical grid scheduling method is designed. First, constraints on the safe and economical operation of renewable energy power systems are defined. Then, the quintuple of Markov decision process is defined under the framework of deep reinforcement learning, and the dispatching optimization problem is transformed into Markov decision process. To solve the problem of low sample data utilization in online reinforcement learning strategies, a PPO optimization algorithm based on the Kullback–Leibler (KL) divergence penalty factor and importance sampling technique is proposed, which transforms on-policy into off-policy and improves sample utilization. Finally, the simulation analysis of the example shows that in a power system with a high proportion of renewable energy generating units connected to the grid, the proposed scheduling strategy can meet the load demand under different load trends. In the dispatch cycle with different renewable energy generation rates, renewable energy can be absorbed to the maximum extent to ensure the safe and economic operation of the grid.

SpacePix2: SOI MAPS detector for space radiation monitoring

The SpacePix2 monolithic active pixel sensor (MAPS) is a novel ASIC for space radiation monitoring designed in a 180 nm SOI CMOS technology. The active sensor area is 3.84 × 3.84 mm2, pixel matrix is arranged as a 64 × 64 array with 60 µm pitch. The pixel front-end amplifier signal range is 2–80 ke−, extended up to 30 Me− using a backside channel. Diodes integrated in the handle wafer in each pixel are biased at −150 V. Impinging particle generates a charge pulse converted to voltage pulse by charge sensitive amplifier (CSA). Maximum voltage memorized by the peak detector hold (PDH) circuit is digitized using on-chip 10-bit asynchronous column SAR ADCs. Two readout interfaces are available, 400 MHz LVDS and 50 MHz SPI. Total power consumption is 43 mA from a 1.8 V power supply. The ASIC has been tested in space in low Earth orbit (LEO) and sample data are shown.

Low-sample classification in NIDS using the EC-GAN method

Numerous advanced methods have been applied throughout the years for the use in Network Intrusion Detection Systems (NIDS). Among these are various Deep Learning models, which have shown great success for attack classification. Nevertheless, false positive rate and detection rate of these systems remains a concern. This is mostly because of the low-sample, imbalanced nature of realistic datasets, which make models challenging to train. Considering this, we applied a novel semi-supervised EC-GAN method for network flow classifi- cation of CIC-IDS-2017 dataset. EC-GAN uses synthetic data to aid the training of a supervised classifier on low-sample data. To achieve this, we modified the original EC-GAN to work with tabular data. In our approach, WCGAN-GP is used for synthetic tabular data generation, while  a simple deep neural network is used for classification. The conditional nature of WCGAN-GP diminishes the class imbalance problem, while GAN itself solves the low-sample problem. This approach was successful in generating believable synthetic data, which was consequently used for training and testing the EC-GAN. To obtain our results, we trained a classifier on progressively smaller versions of the CIC-DIS-2017 dataset, first via a novel EC-GAN method and then in the conventional way, without the help of synthetic data. We then compared these two sets of results with another author’s results using accuracy, false positive rate, detection rate and macro F1 score as metrics. Our results showed that supervised classifier trained with EC-GAN can achieve significant results even when trained on as little as 25% of the original imbalanced dataset.

Non-Intrusive Load Monitoring Based on the Graph Least Squares Reconstruction Method

Monitoring the operating conditions of residential appliances by collecting the total power consumption data of households has a great significance for smart power utilization. In this study, a graph least squares reconstruction approach is proposed. First, the graph signal is constructed using the collected active power consumption data, providing unique appliance signature information. The signal smoothness of the graph and an iterative least squares reconstruction algorithm are utilized for classification. The proposed graph reconstruction method relies on only low-sampling data from installed smart meters. It aims to address some existing problems of event-based NILM methods like measurement noise, indistinguishable load signatures, and inaccurate power reconstruction. Also, extensive training and associated calculations are not required. Simulation results on the REDD benchmark dataset demonstrate that the proposed method outperforms some of the current state-of-the-art techniques.

State of health estimation of lithium-ion batteries based on the regional frequency

The state of health (SOH) estimation of lithium-ion batteries at low data sampling frequencies has important practical significance in engineering application. This paper proposes a new SOH evaluation method for lithium-ion batteries under the framework of probability density function (PDF). The concepts of regional voltage and regional frequency are introduced, and the SOH models are established based on the experimental cycle data of the nickel-cobalt-aluminium (NCA) battery cell and lithium iron phosphate (LFP) battery module. The results show that the SOH is a simple linear function of the regional frequency in the obtained model, and the fitting R -squared ( R 2 ) of two battery SOH models can reach more than 0.99 with appropriate regional voltage. This method does not directly use the characteristic parameters of PDF and is insensitive to the sampling frequency as low as 1/60 Hz. It is effective for accurate online SOH evaluation in occasions where the sampling frequency cannot be too high, such as large energy storage power station. Moreover, this method is more suitable for module SOH evaluation, which can overcome the poor SOH evaluation caused by the inconsistency of cells in a module. • An improved PDF method for SOH evaluation is proposed based on regional frequency. • The regional frequency method can be used at the sampling frequency as low as 1/60 Hz. • The regional frequency method is suitable for SOH modeling of cells and modules. • A good linear positive correlation between battery SOH and regional frequency.

Variation in septic system effluent inputs to tributaries in multiple subwatersheds and approaches to distinguish contributing pathways and areas

Quantifying the contribution of septic systems to contaminant, including nutrient, loading to streams is needed in many watersheds to inform water quality management programs. However, this quantification is challenging due to the distributed locations of septic systems and uncertainties regarding the pathways delivering effluent from septic systems (functioning and failing) to a stream. The objectives of this study were firstly to evaluate how septic effluent inputs to streams vary with stream discharge conditions for multiple subwatersheds with different characteristics (i.e., geology, septic system density, and typical age), and secondly to examine new approaches for distinguishing the pathways and the contributing areas delivering septic effluent to streams. These approaches use the artificial sweetener acesulfame as a conservative tracer for septic effluent in applications of: (i) stream concentration-discharge (C-Q) relationships using low frequency sampling data, (ii) hysteresis behavior in event-based C-Q relationships, and (iii) longitudinal stream sampling. For all nine subwatersheds studied, the amount of septic effluent reaching the subwatershed outlets was considerably higher during high stream discharge (event) conditions compared to low discharge (baseflow) conditions, suggesting pathways other than groundwater may also be important. Generally, the percentage of septic effluent reaching the outlets was less for subwatersheds with newer households compared to those with older households. The combined interpretation of low frequency and event-based C-Q relationships indicate that complex pathways control the delivery of septic effluent to the subwatershed outlets. The interpretations suggest that groundwater pathways may dominate in some subwatersheds, while more rapid pathways associated with failing septic systems (e.g., overland runoff) may be important in others. Finally, longitudinal stream sampling illustrate the potential of acesulfame data to identify key areas contributing septic effluent to the stream. The novel approaches used here can be applied to guide future investigations aiming to quantify and manage water quality impairment from septic systems.

A general framework for supervised structural health monitoring and sensor output validation mitigating data imbalance with generative adversarial networks-generated high-dimensional features

This study proposes a novelty-classification framework that applies to structural health monitoring (SHM) and sensor output validation (SOV) problems. The proposed framework has simple high-dimensional features with several advantages. First, the feature extraction method is extensively applicable to instrumented structures. Second, the high-dimensional features’ utilization alleviates one of the main issues of supervised novelty classifications, namely, imbalanced datasets and low-sampled data classes. Recurrent Neural Networks are employed for the classification of high-dimensional features. Furthermore, generative adversarial networks (GAN) are trained with low-sampled data classes’ high-dimensional features for generating new data objects. The generated data objects are combined with the initial training set for improving classification results. The proposed framework is studied on two SHM and SOV datasets. The SHM dataset has twenty-one data classes, with a total test accuracy of 99.60% compared to another study with 88.13% accuracy. The SOV classification shows improved results with a mean accuracy of 96.5% compared to three other studies with mean accuracy values of 93.5%, 92.97%, and 71.1%. Furthermore, the integration of GAN’s generated data objects with low-sampled classes improved those classes’ mean F1 score from 44.77% to 64.58% and from 73.39% to 90.84% on SOV and SHM case studies, respectively. The integration of GAN-generated data objects with the initial low-sampled data classes for accuracy improvement shows more potential in the SHM dataset than the SOV case, which can be due to the signal pattern-based labeling logic of SOV datasets.

A Maximum Entropy Estimator for the Average Survival Time Differences between Two Groups

Background: Statistical methods commonly used in survival analysis typically provide the probability that the difference between groups is due to chance, but do not offer a reliable estimate of the average survival time difference between groups (the difference between median survival time is usually reported). Objective: We suggest a Maximum-Entropy estimator for the average Survival Time Difference (MESTD) between groups. Methods: The estimator is based on the extra survival time, which should be added to each member of the group, to produce the maximum entropy of the result (resulting in the groups becoming most similar). The estimator is calculated only from time to event data, does not necessarily assume hazard proportionality and provides the magnitude of the clinical differences between the groups. Results: Monte Carlo simulations show that, even at low sample numbers (much lower than the ones needed to prove that the two groups are statistically different), the MESTD estimator is a reliable predictor of the clinical differences between the groups, and therefore can be used to estimate from (low sample numbers) preliminary data whether or not the large sample number experiment is worth pursuing. Conclusion: By providing a reasonable estimate for the efficacy of a treatment (e.g., for cancer) even for low sample data, it might provide useful insight in testing new methods for treatment (for example, for quick testing of multiple combinations of cancer drugs).

Detection and Analysis of Electromechanical Oscillation in Power Systems with Low-Sampled Data Using Modal Analysis Methods

Electromechanical oscillations between interconnected generators are considered a major threat to the secure operation of power systems. Therefore, oscillation monitoring systems in real-time are of critical importance to detect the danger of poorly damped oscillations. For the detection and analysis of the oscillations, high-temporal-resolution measurements are required according to the Nyquist theorem. This paper proposes a novel algorithm for the identification of electromechanical oscillations using low-sampled data such as supervisory control and data acquisition (SCADA) measurements. The lack of temporal resolution of the data is compensated by using low-sampled data sets at multiple different locations. At a target location, a high-sampled data-signal can be reconstructed using mode shape information obtained from model-based modal analysis. The variable projection method is then used to detect oscillations and estimate oscillation components including frequency and damping ratio. Case studies based on practical Korean power systems are presented to evaluate the performance of the proposed method. Simulation results show that the proposed method can detect and identify electromechanical oscillations with low-sampled data.

Smart Energy Meter: Applications, Bibliometric Reviews and Future Research Directions

ABSTRACT Now, Smart Energy Meters (SEMs) have become very popular in controlling and managing electricity consumption in residential, commercial, industrial sectors, to name a few. SEMs collects the fine-grained energy-consumption data at a regular time interval (every hour). Such a low sampling data rate can be useful in research, evaluating the higher consumption of electricity, etc. However, existing research work cannot adequately assess the qualitative, quantitative, and logical properties of the data that could theoretically improve the performance of classification and analytics. The purpose of this bibliometric analysis is to understand the reach of SEMs and its data analysis worldwide and across the various applications to make the concept clear for future researchers. The Scopus, Semantic Scholar and Crossref databases are used for performing the bibliometric analysis. This research study revealed that the insights of the SEMs data-enabled significant advancement in many fields such as energy consumption pattern analysis and prediction, demand response, load profiling, and direct-indirect phone charging analysis.

Positron emission tomography image reconstruction using feature extraction.

To reduce the cost of positron emission tomography (PET) scanning systems, image reconstruction algorithms for low-sampled data have been extensively studied. However, the current method based on total variation (TV) minimization regularization nested in the maximum likelihood-expectation maximization (MLEM) algorithm cannot distinguish true structures from noise resulting losing some fine features in the images. Thus, this work aims to recover fine features lost in the MLEM-TV algorithm from low-sampled data. A feature refinement (FR) approach previously developed for statistical interior computed tomography (CT) reconstruction is applied to PET imaging to recover fine features in this study. The proposed method starts with a constant initial image and the FR step is performed after each MLEM-TV iteration to extract the desired structural information lost during TV minimization. A feature descriptor is specifically designed to distinguish structure from noise and artifacts. A modified steepest descent method is adopted to minimize the objective function. After evaluating the impacts of different patch sizes on the outcome of the presented method, an optimal patch size of 7×7 is selected in this study to balance structure-detection ability and computational efficiency. Applying MLEM-TV-FR algorithm to the simulated brain PET imaging using an emission activity phantom, a standard Shepp-Logan phantom, and mouse results in the increased peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) as comparing to using the conventional MLEM-TV algorithm, as well as the substantial reduction of the used sampling numbers, which improves the computational efficiency. The presented algorithm can achieve image quality superior to that of the MLEM and MLEM-TV approaches in terms of the preservation of fine structure and the suppression of undesired artifacts and noise, indicating its useful potential for low-sampled data in PET imaging.

Temperature prediction based on a space–time regression-kriging model

ABSTRACTMany phenomena exist in the space–time domain, often with a low data sampling rate and sparsely distributed network of observed points. Therefore, spatio-temporal interpolation with high accuracy is necessary. In this paper, a space–time regression-kriging model was introduced and applied to monthly average temperature data. First, a time series decomposition was applied for each station, and a multiple linear regression model was used to fit space–time trends. Second, a valid nonseparable spatio-temporal variogram function was utilized to describe similarities of the residuals in space–time. Finally, space–time kriging was applied to predict monthly air temperature. Jackknife techniques were used to predict the monthly temperature at all stations, with correlation coefficients between predictions and observed data very close to 1. Moreover, to evaluate the advantages of space–time kriging, pure time forecasting also was executed employing an autoregressive integrated moving average (ARIMA) model. The results of these two methods show that both mean absolute error (MAE) and root-mean-square error (RMSE) of space–time prediction are much lower than those of the pure time forecasting. The estimated temperature curves for stations also show that the former present a conspicuous improvement in interpolation accuracy when compared with the latter.

THE IMPLEMENTATION OF DISAGGREGATION ALGORITHM IN THE ANALYSIS OF ENERGY CONSUMPTION BASED ON THE INTERNET-OF- THINGS TECHNOLOGY

The effort to reduce the mass energy usage without involvement of consumers is not effective. Thereby, creating a pathway for anyone of consumer to be much more involved in the energy-saving effort. The implementation of disaggregation algorithm in the analysis of energy consumption is to recognize when and which appliance has the largest energy consumption and being able to control the state of all appliances from anywhere. In this research, the principle of disaggregation is event-based and low-sampling data frequency. A KWH-meter is used to send power data to the cloud server via MQTT protocol. The cloud server gathers the energy-consumption data, analyses them and then disaggregates them. The output of the disaggregation algorithm would tell the state (on/off), the average power and the percentage of energy consumed by each appliances. The output will then be sent from the cloud server to Android Apps via MQTT protocol. Then the consumer can easily access and control the energy consumption from their smartphone after knowing it through the disaggregation algorithm.