Input Time Research Articles

Quantitative improvement of streamflow forecasting accuracy in the Atlantic zones of Canada based on hydro-meteorological signals: A multi-level advanced intelligent expert framework

Developing reliable streamflow forecasting models is critical for hydrological tasks such as improving water resource management, analyzing river patterns, and flood forecasting. In this research, for the first time, an emerging multi-level TOPSIS (technique for order preference by similarity to the ideal solution) -based hybridization comprised of the Boruta classification and regression tree (Boruta-CART) feature selection, multivariate variational mode decomposition (MVMD), and a hybrid Convolutional Neural Network (CNN) Bidirectional Gated Recurrent Unit (CNN-BiGRU) deep learning was adopted to multi-temporal (one and three days ahead) forecast the daily streamflow in the Rivers of Prince Edward Island, Canada. For this aim, in the first step, the Boruta-CART feature selection technique determines the most effective lagged components among all the antecedent two-day information (i.e., t-1 and t-2) of hydro-meteorological features (from 2015 to 2020), including the water level, mean air temperature, heat degree days, total precipitation, dew point temperature, and relative humidity in the Bear and Winter Rivers of Prince Edward Island, Canada. Afterwards, a multivariate variational mode decomposition (MVMD) decomposes the input time series to decrease the complexity and non-linearity of the non-stationary ones before feeding the deep learning (DL) models. Here, the CNN-GRU was employed as the primary DL model, along with the kernel extreme machine method (KELM), random variational function link (RVFL), and hybrid CNN bidirectional recurrent neural network (CNN-BiRNN) as the comparative models. A TOPSIS scheme applying several performance measures like the correlation coefficient (R), root mean square error (RMSE), and reliability was designed for the robustness assessment of the hybrid (MVM-CNN-BiGRU, MVM-CNN-BiRNN, MVM-RVFL, and MVM-KELM) and standalone models. The computational outcomes revealed that in the Bear River, the MVM-CNN-BiGRU, owing to its best forecasting performance (one day ahead: TOPSIS score 1, R = 0.960, RMSE = 0.098, and reliability = 65.082; three days ahead: TOPSIS score = 0.999, R = 0.924, and RMSE = 0.33) outperformed the other hybrid models, followed by the MVM-CNN-BiRNN, MVM-RVFL, and MVM-KELM, respectively. Moreover, in the Winter River, the MVM-CNN-BiGRU in terms of (one-day ahead: TOPSIS score = 0.890, R = 0.955, RMSE = 0.274, and reliability = 34.004; three-days ahead: TOPSIS score = 0.686, R = 0.924, and RMSE = 0.330) was superior to the other models. The provided expert system could be vital in the local flood decision-making process, in the absence of streamflow information as input modeling, during the flood seasons to reduce flood damage in residential areas.

METHOD FOR AGENT-ORIENTED TRAFFIC PREDICTION UNDER DATA AND RESOURCE CONSTRAINTS

Context. Problem of traffic prediction in a city is closely connected to the tasks of transportations in a city as well as air pollution detection in a city. Modern prediction models have redundant complexity when used for separate stations, require large number of measuring stations, long measurement period when predictions are made hourly. Therefore, there is a lack of method to overcome these constraints. The object of the study is a city traffic.
 Objective. The objective of the study is to develop a method for traffic prediction, providing models for traffic quantification at measuring stations in the future under data and resource constraints.
 Method. The method for agent-oriented traffic prediction under data and resource constraints was proposed in the paper. This method uses biLSTM models with input features, including traffic data obtained from agent, representing target station, and other agents, representing informative city stations. These agents are selected by ensembles of decision trees using Random Forest method. Input time period length is proposed to set using autocorrelation data.
 Results. Experimental investigation was conducted on traffic data taken in Madrid from 59 measuring stations. Models created by the proposed method had higher prediction accuracy with lower values of MSE, MAE, RMSE and higher informativeness compared to base LSTM models.
 Conclusions. Obtained models as study results have optimal number of input features compared to the known models, do not require complete system of city stations for all roads. It enables to apply these models under city traffic data and resource constraints. The proposed solutions provide high informativeness of obtained models with practically applicable accuracy level.

H∞Control of the Actively Controlled Systems With Input Time Delay Based on CHsCTA Model Transformation

H_∞Control of the Actively Controlled Systems With Input Time Delay Based on CHsCTA Model Transformation

Model Predictive Control of Deep Neural Network Model with ReLU Structure using Mixed Integer Programming

This study proposes a model predictive control framework using a mixed integer programming for the deep neural network (DNN) model with rectified linear unit (ReLU). The accurate DNN model with ReLU using the past input time series for the diesel engine is constructed. In this method, the ReLU is represented by inequalities including 0-1 variables (binaries). The model predictive control problem is formulated as a mixed integer programming problem and is applied to the control problem of the diesel engine. Dynamical DNN model of a diesel engine is constructed using the experimental data. This model can represent complicate engine phenomena such as strong nonlinearity and delay. Simulation results of the worldwide harmonized transient cycle (WHTC) verify the effectiveness of the proposed method. The difficulties of non-linearity, delay, and interference in the control problem of the diesel engine are solved by using a systematic model predictive control of the DNN model.

Procjena mogućnosti priključenja fotonaponskog sistema na srednjenaponsku distributivnu mrežu primjenom sekvencijalne Monte Carlo simulacije

In this paper, a simulator for assessing the possibility of PV integration to the medium voltage distribution network is developed. The PV location and peak power are defined in advance. The simulator is based on sequential Monte Carlo simulation with various possible uncertainties such as photovoltaic production and load. In addition, the sequential simulation enables detection of period when the network constraints are violated. For this analysis, time series of essential variables such as PV system production and load are modeled. Based on the results of power flow calculation, probability distributions of the voltage at various nodes, the loads on individual lines as well as the probability of network constraints exceeding are determined. Depending on the network constraints, possibility of connecting the PV system to the network at the desired location is assessed. The power flow calculation is executed by using the OpenDSS simulator, while the input time series are prepared in Matlab. The method presented in this paper is tested on IEEE 33 distribution test network.

Stratigraphic evolution and sediment dispersal in a marine rift-basin: The case of the Lewis trough, Eastern Dampier sub-basin (Australia)

The understanding of the stratigraphic evolution of marine rift basins is important for both “pure” scientific and economic reasons. From a hydrocarbon exploration point of view the main goal is to develop tools to predict reservoir location, quality, and extent. Stratigraphic models try to do that, and only historical cases can be used to test those models. On the other hand, depositional model cannot precisely predict reservoir location as factors unique to the specific basin under analysis affect sediment distribution (basin physiography, type of flows, among others). This aspect can be assessed only on a case-by-case basis and can be used for initial hypothesis generation based on analogues. This paper shows a case history study based on the integration of well logs, and 3D seismic data across the Eastern Dampier sub-basin, Northwester Shelf of Australia. Advanced seismic interpretation techniques, calibrated to well data, enable the definition of a chrono-stratigraphic framework of the late syn-rift sequence (Late Jurassic) of the Lewis Trough, and the description of the sediment distribution through the interpretation of geomorphological features. During the Late Jurassic three different channel systems deliver sandstones to the deep-water part of the basin through multiple entry points. The timing of the sediment input seems to be linked to variation of sediment input as during that time relative sea-level was slowly rising. The location of the entry points is related to local physiography and defies any attempt of prediction based on general “rules.” From a stratigraphic point view, the findings of the study support the idea that, in rift basin, mechanical subsidence is the primary control for accommodation space, with episodic pulses of extension followed by longer periods of tectonic quiescence. It is during these episodic movements along the basin's bounding faults that space for sedimentation is created. This interpretation is in striking contrast with the basic assumptions of the conventional sequence stratigraphy model.

Method of Synthesis of Devices for Parallel Stream Calculation of Scalar Product in Real Time

A graph scheme of a generalized algorithm for parallel stream calculation of the scalar product was developed. The proposed algorithm uses the same type of operations for forming a partial product that is calculated starting from the lowest digits of the multipliers. The developed algorithm of parallel stream calculation of the scalar product is performed with the use of operations for forming partial products, calculating the macro-partial product, and adding it to the partial result shifted to the right by the number of digits that were used in the formation of partial products. It is suggested that the development of FPGA structures of devices for parallel stream calculation of the scalar product be carried out according to the following principles: use of the same type of conveyor steps; performing calculations based on addition, inversion, and shift operations; performing the calculation of the scalar product as a single operation; regularity and localization of connections between conveyor steps; coordination of the duration of the conveyor time with the time of data input and the time of output of calculation results; space-time parallelization of the process of calculating the scalar product. The algorithm and structure of the parallel stream device for calculating the scalar product with direct formation of partial products based on the analysis of one order of multipliers, which ensures operation with the smallest conveyor cycle, has been developed. The algorithm and structure of the parallel stream device for calculating the scalar product with the formation of partial products for the sum of two pairs of products with the analysis of one order of multipliers, which is advisable to use for a small number of operands, have been developed. The algorithm and structure of a parallel stream device for calculating the scalar product with the formation of partial products according to the modified Booth algorithm have been developed, which ensures a reduction in equipment costs when processing operands with n≥24 bits. The algorithm and structure of the device for calculating the scalar product with the formation of group partial products have been developed, which provides the lowest equipment costs in the case of n=8 for N>8. A method for the synthesis of FPGA devices for parallel stream calculation of the scalar product in real-time has been developed. The proposed method ensures high efficiency of the use of the equipment due to the selection of the algorithm for the formation of partial products and the structure of the device from the list of developed ones and the coordination of the cycle of the conveyor of the selected structure with the time of arrival of input data.

Adaptive dynamic event-triggered cluster consensus for second-order delayed MASs with input saturation

In this paper, the cluster consensus problem of second-order multi-agent systems (SOMASs) is investigated in which cooperative-competitive interactions among agents, input time delay, and input saturation are involved simultaneously. A novel adaptive dynamic event-triggered control (ETC) protocol is first designed to realize cluster consensus. Since the constant in static trigger condition is replaced by a dynamic parameter, the new control scheme can reduce the trigger times and save energy consumption. Then, switching topology problem of heterogeneous multi-agent systems (MASs) is also discussed based on the dynamic ETC strategy. Some sufficient conditions for achieving the cluster consensus of SOMASs are proposed by utilizing the Lyapunov stability theory and algebraic graph theory. Especially, to solve the saturation problem, Lagrange mean value theorem is adopted. Moreover, it is also proved that no Zeno behavior occurs under the proposed control protocol. Finally, two examples justify the significance of our theoretical results.

NEURAL ORDINARY DIFFERENTIAL EQUATIONS FOR TIME SERIES RECONSTRUCTION

Context. Neural Ordinary Differential Equations is a deep neural networks family that leverage numerical methods approaches for solving the problem of time series reconstruction, given small amount of unevenly distributed samples.
 Objective. The goal of the following research is the synthesis of a deep neural network that is able to solve input signal reconstruction and time series extrapolation task.
 Method. The proposed method exhibits the benefits of solving time series extrapolation task over forecasting one. A model that implements encoder-decoder architecture with differential equation solving in latent space, is proposed. The latter approach was proven to demonstrate outstanding performance in solving time series reconstruction task given a small percentage of noisy and uneven distributed input signals. The proposed Latent Ordinary Differential Equations Variational Autoencoder (LODE-VAE) model was benchmarked on synthetic non-stationary data with added white noise and randomly sampled with random intervals between each signal.
 Results. The proposed method was implemented via deep neural network to solve time series extrapolation task.
 Conclusions. The conducted experiments have confirmed that proposed model solves the given task effectively and is recommended to apply it to solving real-world problems that require reconstructing dynamics of non-stationary processes. The prospects for further research may include the process of computational optimization of proposed models, as well as conducting additional experiments involving different baselines, e. g. Generative Adversarial Networks or attention Networks.

Optimization of Real-Time Operational Data Monitoring Through Installing The Device on the Unit

Coal is the largest natural resource in Indonesia and has a vital role in global energy in various industries. PT Putra Perkasa Abadi (PPA), a mining contractor in South Kalimantan, faces problems with manual reporting being late and prone to errors. This project aims to speed up the digitization of the reporting process by providing PPA Team devices in the Heavy Equipment (A2B) and Support units. This is in line with the commitment of the PPA Mining Work Safety, Environment and Quality (K3PLM) Policy. This implementation starts from observation and data collection by selecting SS6 devices and creating designs for each unit. Continue with device installation, analyze problems, and repair installation. Following, carry out trials and socialize the use of the device to operators. With the implementation of the PPA Team application on the SS6 Monitor, Operators at PPA can independently input data on operational activities, including check-in / check-out units. It will reduce paper in operational reporting by providing an SS6 Monitor device. In addition, cost calculations can be adjusted to actual data, and operators can monitor unit Hour Meter (HM) data and operational activities directly on the SS6 Monitor. This reduces the duration of data input time, but also reduces the intensity of crowding on communication radios, and minimizes data input errors. Therefore, optimizing operational monitoring at PPA by installing devices on units can increase efficiency, including HM input for cost calculations, efficient unit HM reporting, fast check-in / check-out processes, digital unit breakdown reporting for more efficient communication, rapid digital validation of P2H, and moving from paper forms to digital tools for more sustainable work practices.

TS2ACT

Human Activity Recognition (HAR) based on embedded sensor data has become a popular research topic in ubiquitous computing, which has a wide range of practical applications in various fields such as human-computer interaction, healthcare, and motion tracking. Due to the difficulties of annotating sensing data, unsupervised and semi-supervised HAR methods are extensively studied, but their performance gap to the fully-supervised methods is notable. In this paper, we proposed a novel cross-modal co-learning approach called TS2ACT to achieve few-shot HAR. It introduces a cross-modal dataset augmentation method that uses the semantic-rich label text to search for human activity images to form an augmented dataset consisting of partially-labeled time series and fully-labeled images. Then it adopts a pre-trained CLIP image encoder to jointly train with a time series encoder using contrastive learning, where the time series and images are brought closer in feature space if they belong to the same activity class. For inference, the feature extracted from the input time series is compared with the embedding of a pre-trained CLIP text encoder using prompt learning, and the best match is output as the HAR classification results. We conducted extensive experiments on four public datasets to evaluate the performance of the proposed method. The numerical results show that TS2ACT significantly outperforms the state-of-the-art HAR methods, and it achieves performance close to or better than the fully supervised methods even using as few as 1% labeled data for model training. The source codes of TS2ACT are publicly available on GitHub1.

Temporal collaborative attention for wind power forecasting

Wind power serves as a clean and sustainable form of energy. However, its generation is fraught with variability and uncertainty, owing to the stochastic and dynamic characteristics of wind. Accurate forecasting of wind power is indispensable for the efficient planning, operation, and grid integration of wind energy systems. In this paper, we introduce a novel forecasting method termed Temporal Collaborative Attention (TCOAT). This data-driven approach is designed to capture both temporal and spatial dependencies in wind power generation data, as well as discern long-term and short-term patterns. Utilizing attention mechanisms, TCOAT dynamically adjusts the weights of each input variable and time step based on their contextual relevance for forecasting. Furthermore, the method employs collaborative attention units to assimilate directional and global information from the input data. It also explicitly models the interactions and correlations among different variables or time steps through the use of self-attention and cross-attention mechanisms. To integrate long-term and short-term information effectively, TCOAT incorporates a temporal fusion layer that employs concatenation and mapping operations, along with hierarchical feature extraction and aggregation. We validate the efficacy of TCOAT through extensive experiments on a real-world wind power generation dataset from Greece and compare its performance against twenty-two state-of-the-art methods. Experimental results demonstrate that TCOAT outperforms existing methods in terms of both accuracy and robustness in wind power forecasting. Moreover, we conduct a generality study on an additional real-world dataset from a different climate condition and wind power characteristics. The results show that TCOAT can achieve comparable or better performance than the state-of-the-art methods, confirming the generalization ability of TCOAT.

Linear and Non-Linear Spatio-Temporal Input Selection In Wireless Traffic Networks Prediction using Recurrent Neural Networks

For the optimization of computer networks with high bandwidth requirements, it is necessary to predict the traffic of the wireless network. Its goal is to reduce maintenance costs and improve internet services. Feature selection is a major issue in multivariate time series (MTS) spatio-temporal modeling. Another problem is the dependency between input features, time lags, and spatial factors, so an appropriate model is needed. This study aims to provide solutions to two problems. The first is to improve a feature extraction and selection process in spatio-temporal MTS data for relevant features using Detrended Partial Cross-Correlation Analysis (DPPCA) and nonredundant features associated with linear using Pearson's correlation (PC) filters and non-linear associations using Symmetrical Uncertainty (SU) and a combination of both PCSUF. The second is to develop a spatiotemporal framework model using recurrent neural networks (RNNs) to get better performance than the traditional model. These methods are combined and tested using a data set of cellular networks with one hour intervals during November in three locations. Testing the effectiveness of the feature selection technique showed that 27.6% of the total extracted features were. The forecasting model with the DPCCA-SU-RNN combination method is the best performance by having RMSE = 380.7, R2 = 97% and MAPE = 10%.

Optimized derivative fast Fourier transform with high resolution and low noise from encoded time signals: Ovarian NMR spectroscopy

The unfiltered derivative fast Fourier transform (dFFT) of degrees higher than two fails flagrantly for encoded time signals. These data are always dominated by noise at larger times of encodings. Such a breakdown is due to processing the unweighted product of the time signal and the time power function. The latter is generated by the frequency derivative operator applied to the fast Fourier transform (FFT). As a result, the unfiltered dFFT cannot separate the overlapped resonances and it dramatically decreases signal-to-noise ratio (SNR) relative to the FFT. This problem is solved by a derivative-adapted optimization with the properly attenuated filters. The ensuing optimized dFFT achieves the long sought simultaneous enhancement of both resolution and SNR. It uncovers the genuine resonances hidden within overlapping peaks to enable quantitative interpretations. It does not impose any model on the input time signals nor on the output lineshape in the spectra. It is computationally expedient as it uses the Cooley-Tukey fast algorithm. The present applications deal with time signals encoded by in vitro NMR spectroscopy from human malignant ovarian cyst fluid. A remarkably successful performance of the optimized dFFT is demonstrated for reconstructed spectra of potentially added value in clinical decision-making.

Measurement of multidimensional reconstruction fund planning on the effectiveness of urban post-disaster recovery and reconstruction using indirect economic impact

An increasing number of record-breaking extreme precipitation events worldwide has had an incalculable socioeconomic impact, challenging the ability of cities to recover and reconstruct after disasters. The key to the effectiveness of post-disaster recovery and reconstruction lies in the rational investment and allocation of reconstruction funds. However, few studies have examined the effectiveness of reconstruction fund planning in improving urban recovery and reconstruction from a post-disaster indirect economic loss assessment perspective. This study uses an improved input-output model to evaluate the degree of change in indirect economic losses under multidimensional reconstruction fund planning (funds' amount, the time efficiency of investment, and funds allocation) to measure the effectiveness. We exemplify this through a case study: a record-breaking extreme urban flood and its reconstruction planning occurred in Henan Province, China, in July 2021. The results indicate that (1) under the premise of ensuring sufficient financial input, improving the input time efficiency is a decision with a higher benefit-cost ratio, which can improve the effectiveness by 140%. (2) If funds are guaranteed, improving time efficiency is better than adding extra funds, and when both reach 80%, additional funds may not further improve the effectiveness. (3) The funds need to meet at least the recovery production needs (about 45% of direct economic losses); otherwise, the funds cannot be used to improve fortifications. The study results will provide theoretical and practical support for urban decision-makers to develop efficient and rational recovery and reconstruction programs.

Spiking ink drop spread clustering algorithm and its memristor crossbar conceptual hardware design

<span lang="EN-US">In this study, a novel neuro-fuzzy clustering algorithm is proposed based on spiking neural network and ink drop spread (IDS) concepts. The proposed structure is a one-layer artificial neural network with leaky integrate and fire (LIF) neurons. The structure implements the IDS algorithm as a fuzzy concept. Each training data will result in firing the corresponding input neuron and its neighboring neurons. A synchronous time coding algorithm is used to manage input and output neurons firing time. For an input data, one or several output neurons of the network will fire; confidence degree of the network to outputs is defined as the relative delay of the firing times with respect to the synchronous pulse. A memristor crossbar-based hardware is utilized for hardware implementation of the proposed algorithm. The simulation result corroborates that the proposed algorithm can be used as a neuro-fuzzy clustering and vector quantization algorithm.</span>

Wavelet transform and deep learning-based obstructive sleep apnea detection from single-lead ECG signals.

Sleep apnea is a common sleep disorder. Traditional testing and diagnosis heavily rely on the expertise of physicians, as well as analysis and statistical interpretation of extensive sleep testing data, resulting in time-consuming and labor-intensive processes. To address the problems of complex feature extraction, data imbalance, and low model capacity, we proposed an automatic sleep apnea classification model (CA-EfficientNet) based on the wavelet transform, a lightweight neural network, and a coordinated attention mechanism. The signal is converted into a time-frequency image by wavelet transform and put into the proposed model for classification. The effects of input time window, wavelet transform type and data balancing on the classification performance are considered, and a cost-sensitive algorithm is introduced to more accurately distinguish between normal and abnormal breathing events. PhysioNet apnea ECG database was used for training and evaluation. The 3-min Frequency B-Spline wavelets transform of ECG signal was carried out, and Dice Loss was used to train the classification model of sleep breathing. The classification accuracy was 93.44%, sensitivity was 88.9%, specificity was 96.2% and most indexes were better than other related work.

Application of the Single-Stage Process of Dark Fermentation and Microbial Electrolysis to Improve Hydrogen Productivity from Water Hyacinth

The production of hydrogen (H2) from water hyacinth (WH) can contribute to reducing both the negative impact of WH on ecosystems and dependence on fossil fuels. In this study, the combination of dark fermentation (DF) and microbial electrolysis cell (MEC) in a single reactor, namely, sDFMEC, was investigated to improve the H2 productivity of WH. Furthermore, the intermittently applied voltage (I-Eapp) scheme and various methane (CH4) inhibition methods, including air exposure, heat treatment, and chloroform (CHCl3) addition, were applied for performance enhancement purposes. The findings indicated that with a sufficient duty time of external energy input (less than 1 hour), the intermittent mode can enhance the performance of WH-fed sDFMEC but does not significantly inhibit CH4 formation. While air exposure and heat treatment damaged both methanogens and exoelectrogens, lowering sDFMEC performance, additional CHCl3 showed the best selective and long-term inhibition on methanogens (over 350 operation hours without further addition). Overall, the combination of the I-Eapp scheme and CHCl3 applied in WH-fed sDFMEC achieved a yield of 670.1 ± 15.2 mL − H 2 / g − VS , around 160% higher than the normal condition.

SolBeePop: A model of solitary bee populations in agricultural landscapes

Abstract With their large diversity of species, solitary bees are important pollinators of crops and native plant communities in agricultural landscapes. Stressors such as changing landscapes, climate and pesticide exposures may affect populations differently, dependent on each species' ecological traits. We developed a population model for solitary bees, SolBeePop, which can be applied to simulate a variety of species by using species‐specific traits, including the nesting strategies. Species' phenological traits are mechanistically combined with input time series capturing temporal and spatial variability in landscape compositions. Calibration and validation of the model with empirical study data demonstrate that the model can capture realistic dynamics in bee populations. In simulations conducted representing four species, Osmia bicornis, Megachile rotundata, Nomia melanderi and Eucera pruinosa, identical conditions and assumed nesting resource limitations resulted in different population‐level outcomes, indicating the importance of interactions between external factors and species‐specific traits including phenological, survival and reproductive traits. Synthesis and applications. The publicly available model is intended as a tool for the assessment of population‐level outcomes of stressors, for instance, the limitation of floral resources in agricultural landscapes, limitation of nesting habitat and the exposure to pesticides. Realistic landscape scenarios can be tested and available data for one species can be used to estimate outcomes in other solitary bee species, informing conservation plans and risk assessment approaches to support managed and natural populations in the field.

Nonreciprocal propagation of optical pulses in a one-dimensional photonic crystal with two Weyl semimetal-based defects

We investigated the nonreciprocal propagation properties of an optical pulse in a one-dimensional photonic crystal containing Weyl semimetal-based defect layers. High optical nonreciprocity, without the need for high-intensity input light sources or external magnetic fields, is the most essential property of the Weyl semimetal. The transfer matrix and Fourier transform techniques are used to obtain the time envelopes of the transmitted pulses. It is demonstrated that the transmitted pulse may attain two different time envelopes by reversing the impinging direction of the input pulse. Then, the effects of the input pulse carrier frequency and time width on the characteristics of the transmitted pulse are presented and shown that reversing the impinging direction of the input pulse has a significant effect on the length and energy of the transmitted pulse. The results may have applications in designing pulse-shaping filters and transferring digital data.