Lower Average Error Research Articles

Comparison of non-linear time series models (Beta-t-EGARCH and NARMAX models) with Radial Basis Function Neural Network using Real Data

This paper presents a comparison of three different non-linear time series modelling approaches: NARMAX (Non-linear Autoregressive Moving Average with Exogenous Inputs), Beta-t-EGARCH (Beta t Exponential Generalized Autoregressive Conditional Heteroscedasticity), and Radial Basis Function Neural Networks (RBFNN) applied to weekly stock market index data. We will explain three types of models and compare their compositions and structures. Then, we will show which model gives better predictions. To study series data, the comparison involved analysing the structure of the model and its errors in various time series models and summarising their findings. We divide the data into two parts: training data to structure the time series and testing. The training data tests the model's predictions. Then, we can analyse the model with the errors and the best deterrence predictions. After selecting the NARMAX and Beta-t-EGARCH models, we test them with specific criteria. The best choice is finding the model with the lowest average errors. For this study, we analysed the weekly average closing of the Aramco 2222 index from 15 December 2019 to 16 July 2023 and made 187 observations.

Video Abnormal Behavior Recognition and Trajectory Prediction Based on Lightweight Skeleton Feature Extraction.

Video action recognition based on skeleton nodes is a highlighted issue in the computer vision field. In real application scenarios, the large number of skeleton nodes and behavior occlusion problems between individuals seriously affect recognition speed and accuracy. Therefore, we proposed a lightweight multi-stream feature cross-fusion (L-MSFCF) model to recognize abnormal behaviors such as fighting, vicious kicking, climbing over the wall, et al., which could obviously improve recognition speed based on lightweight skeleton node calculation, and improve recognition accuracy based on occluded skeleton node prediction analysis in order to effectively solve the behavior occlusion problem. The experiments show that our proposed All-MSFCF model has a video action recognition average accuracy rate of 92.7% for eight kinds of abnormal behavior recognition. Although our proposed lightweight L-MSFCF model has an 87.3% average accuracy rate, its average recognition speed is 62.7% higher than the full-skeleton recognition model, which is more suitable for solving real-time tracing problems. Moreover, our proposed Trajectory Prediction Tracking (TPT) model could real-time predict the moving positions based on the dynamically selected core skeleton node calculation, especially for the short-term prediction within 15 frames and 30 frames that have lower average loss errors.

SANGRIA: Stacked Autoencoder Neural Networks With Gradient Boosting for Indoor Localization

Indoor localization is a critical task in many embedded applications, such as asset tracking, emergency response, and realtime navigation. In this article, we propose a novel fingerprintingbased framework for indoor localization called SANGRIA that uses stacked autoencoder neural networks with gradient boosted trees. Our approach is designed to overcome the device heterogeneity challenge that can create uncertainty in wireless signal measurements across embedded devices used for localization. We compare SANGRIA to several state-of-the-art frameworks and demonstrate 42.96% lower average localization error across diverse indoor locales and heterogeneous devices.

Additional study of failure behavior for ship propeller: The comparative analysis of life evaluation approaches between BP neural network, LSTM and TV-HSMM

The study focuses on investigating the appropriate method for evaluating the residual life of a propeller and analyzing its fracture characteristics. A fluid-structure interaction simulation is conducted to qualitatively analyze the stress distribution of the propeller during operation, identifying regions prone to stress concentration. Subsequently, fracture surface characteristics are examined, revealing quasi-cleavage fracture features such as tongue striations, dimples, and tearing ridges in the initiation and propagation regions. Additionally, composite fracture characteristics and geometric mutations are observed on the edge of the fracture surface. Three residual life evaluation methods are then proposed, utilizing BP neural network, LSTM, and TV-HSMM. Comparative analysis suggests that the TV-HSMM method offers the advantage of minimal error in residual life evaluation. However, the models based on BP neural network and LSTM are preferred for their lower average error and distribution trends that closely resemble actual life data.

Optimization on the peroxide ratio control strategy of PEMFC system based on immune algorithm

The air supply system is an essential component of a proton exchange membrane fuel cell (PEMFC), and proper control of the oxygen stoichiometry ratio(λo2) is crucial to prevent oxygen starvation in the PEMFC while maintaining efficient and stable performance output. In response to the issues of poor dynamic performance and slow response in traditional λo2 control strategies, this study was conducted to establish a dynamic oxygen flow model of the PEMFC and proposed a λo2 control strategy based on an immune algorithm-fuzzy PID (IAFPID). The performance of the IAFPID control strategy in λo2 control was analyzed under different load requirements, simulated vehicle dynamic loads, and temperature disturbances. The results demonstrate that the IAFPID control strategy has advantages such as fast convergence speed, strong robustness, and good dynamic performance. The efficiency of achieving the optimal stoichiometry ratio for high power output in PEMFC under disturbances was nearly 50% higher with IAFPID control compared to traditional PID, Fuzzy-PID, Genetic Algorithm PID, and Immune Algorithm PID strategies. Furthermore, this method ensures a lower average error in maintaining stable λo2 under dynamic disturbances compared to the IAPID and PID control strategies, with reductions of 12.93% and 37.75%, respectively.

Wi-Cyclops: Room-Scale WiFi Sensing System for Respiration Detection Based on Single-Antenna

Recent years have witnessed the emerging development of single-antenna wireless respiration detection that can be integrated into IoT devices with a single transceiver chain. However, existing single-antenna-based solutions are all limited by the short sensing range within 2-4 m due to noise interference, which makes them difficult to be adopted in most room-scale scenarios. To deal with this dilemma, we propose a room-scale, noise-resistance, and accurate respiration monitoring system, named Wi-Cyclops , 1 which captures CSI changes induced by respiratory movements only via one antenna on commercial WiFi devices. To push the limits of effective sensing distance, we innovatively supply a new perspective to review the CSI samples along the sub-carrier dimension. From this dimension, we find that the interrelationship between sub-carriers with different timestamps still shows a high correlation even though the SNR decreases. Based on that, we analyze the noise characteristics along the sub-carrier dimension and correspondingly design a series of denoising schemes. Specifically, we carefully design a PCA-based denoising method to filter out ambient noises. After that, considering the low distribution densities of the AGC-induced noise, we then remove it by optimizing the DBSCAN denoising method with the K-Means-based adaptive radius search. Extensive experiments demonstrate that our system can work effectively in three typical family scenarios. Wi-Cyclops can achieve 98% accuracy even when the person is 7 m away from the transceiver pair. Compared with the start-of-art single-antenna-based approaches in real scenarios, Wi-Cyclops can improve the sensing range from 3 m to 7 m, which can meet the requirements of room-scale respiration monitoring. Additionally, to show the high compatibility with smart home devices, Wi-Cyclops is deployed on seven commercial IoT devices and still achieves a low average absolute error with 0.41 bpm.

Estimation of greenhouse gas emission due to open burning of rice straw using Sentinel data

In recent decades, Vietnam has gradually become a critical global rice producer. During that production process, residual straw becomes an environmental pollutant due to open burning, raising greenhouse gas emissions. This study combines the optical images of the Sentinel-2 satellite and the radar images of the Sentinel-1 satellite to estimate the dry biomass of rice and to determine gas emissions due to rice straw burning over the fields in Quoc Oai district, Hanoi city for urban environmental management purposes. Sentinel-2 images have been classified into the land covers, thereby identifying the areas of ​​rice cultivation and the areas of ​​burned straw. Meanwhile, the Sentinel-1 radar image has been used to calculate the dry biomass of rice due to its ability to penetrate clouds, an obstacle to optical images in tropical regions. Furthermore, a field trip during harvesting season allows us to measure aboveground dry biomass. Then, the analysis shows a high correlation between the backscatter V.V. and V.H. of the radar image and the in-situ dry biomass (R=0.923 and R2=0.852), with a relatively low average error (RMSE = 6.58 kg/100 m2). By linear regression method, the study found the total rice dry biomass of 28728.5 tons, which was obtained after the Summer rice crop 2020 for the whole Quoc Oai district, of which 2037.91 tons of rice straw have been burned, releasing a large amount of greenhouse gas emission with 2398.6 tons of CO2, 189.5 tons of CO, 18.8673 tons of PM10 dust, 17.2087 tons of PM2.5 dust and some other gases. The identical procedure has also been applied to the western region of Hanoi city center to estimate the amount of gas emissions. This study has proven the effectiveness of an approach and contributed to supporting urban managers in proposing appropriate policies to monitor and protect the environment.

High-temperature mechanical properties of additively manufactured 420 stainless steel

Martensitic stainless steels are indispensable alloys in various high stress and temperature applications such as plastic injection molds and components in steam generators. Subtractive manufacturing methods used to fabricate these parts, however, limits its functionality and performance due to design constraint of cooling channels. This limitation can be resolved by means of additive manufacturing while ensuring that acceptable high-temperature properties can be achieved. In this work, the mechanical behavior of additively manufactured 420 stainless steel (AM420SS) is explored through material constitutive modeling to determine the mathematical model that best describes its flow stress in extreme conditions. This is accomplished by subjecting the samples to hot compression under the strain rates of 0.1–1.0 s−1, and temperatures between 973–1423 K (700 °C–1150 °C) via Gleeble thermomechanical test. The experimental data were used to generate the predictive flow stress curves of constitutive models which includes Johnson-Cook, Zerilli-Armstrong, Zener-Hollomon, and Hensel-Spittel equations. Results showed that Zener-Hollomon and Hensel-Spittel models are the most accurate material constitutive equations with relatively high R values of 0.986 and 0.976, and low average absolute relative error values of 6.96% and 7.69%, respectively. The material constants derived from these models can be applied in finite element analysis simulations to assess the performance of using AM420SS parts at high temperature and strain conditions.

Evaluation of the perturbed hard-sphere-chain equation of state for calculations of methane hydrate formation condition in the presence of ionic liquids

This work aims to evaluate the performance of the Perturbed Hard-Sphere-Chain (PHSC) equation of state (EoS) for the calculation of methane hydrate formation in the presence of ionic liquids (ILs). The van der Waals and Platteeuw model is utilized to calculate the hydrate dissociation pressure. The performance of the PHSC model is compared to cubic EoSs and the Perturbed Chain Statistical Associating Fluid Theory (PC-SAFT) model. Two activity coefficient models are used to estimate the water activity in the presence of ILs. The results show that PHSC EoS with an activity coefficient model provides lower average error values (lower than 2%) and changing the activity coefficient models does not have a great effect on reducing the error. Also, the results of cubic-based EoSs are quite remarkable considering their simplicity. The PHSC model can be utilized as a robust and efficient thermodynamic model for hydrate dissociation pressure calculation in the presence of IL.

Cumulative error correction of inertial navigation systems using LIDAR sensors and extended Kalman filter

Autonomous robots have gained significant attention in research due to their ability to facilitate human work. Navigation systems, particularly localization, present a challenge in autonomous robots. The inertial navigation system is a localization system that uses inertial sensors and a wheel odometer to estimate the robot’s relative position to the initial position. However, the system is susceptible to continuous error accumulation over time due to factors like sensor noise and wheel slip. To address these issues, external sensors are required to measure the robot’s position in the environment. The extended Kalman filter (EKF) method is utilized to estimate the robot’s position based on wheel odometer and light detection and ranging (LIDAR) sensor measurements. In the prediction stage, the input to the EKF is the position measurement from the wheel odometer, while the LIDAR sensor’s position measurement is used in the update stage to improve the prediction stage results. The test results reveal that the EKF’s estimated position has a lower average error compared to the position measurement using the wheel odometer. Therefore, it can be concluded that the EKF technique is effectively applied to the robot and can correct the wheel odometer's cumulative error with the assistance of the LIDAR sensor.

SCRP-Radar: Space-Aware Coordinate Representation for Human Pose Estimation Based on SISO UWB Radar

Human pose estimation (HPE) is an integral component of numerous applications ranging from healthcare monitoring to human-computer interaction, traditionally relying on vision-based systems. These systems, however, face challenges such as privacy concerns and dependency on lighting conditions. As an alternative, short-range radar technology offers a non-invasive, lighting-insensitive solution that preserves user privacy. This paper presents a novel radar-based framework for HPE, SCRP-Radar (space-aware coordinate representation for human pose estimation using single-input single-output (SISO) ultra-wideband (UWB) radar). The methodology begins with clutter suppression and denoising techniques to enhance the quality of radar echo signals, followed by the construction of a micro-Doppler (MD) matrix from these refined signals. This matrix is segmented into bins to extract distinctive features that are critical for pose estimation. The SCRP-Radar leverages the Hrnet and LiteHrnet networks, incorporating space-aware coordinate representation to reconstruct 2D human poses with high precision. Our method redefines HPE as dual classification tasks for vertical and horizontal coordinates, which is a significant departure from existing methods such as RF-Pose, RF-Pose 3D, UWB-Pose, and RadarFormer. Extensive experimental evaluations demonstrate that SCRP-Radar significantly surpasses these methods in accuracy and robustness, consistently exhibiting lower average error rates, achieving less than 40 mm across 17 skeletal key-points. This innovative approach not only enhances the precision of radar-based HPE but also sets a new benchmark for future research and application, particularly in sectors that benefit from accurate and privacy-preserving monitoring technologies.

A deep learning framework for prediction of crop yield in Australia under the impact of climate change

Accurate prediction of crop yields is essential to ensure food security. In this study, a new deep neural networks framework is developed to predict crop yields in Australia, considering the impact of climate change, fertilizer use, and crop area. It is implemented for oats, corn, rice, and wheat crops, and its forecasting performance is benchmarked against five statistical and machine learning methods. All the software codes for the implementation of the proposed framework are freely available. The proposed framework shows the highest forecasting performance for all the considered crop types. It provides 23%, 38%, 39%, and 40% lower average mean absolute error than the benchmark methods for oat, corn, rice, and wheat crops, respectively. The reductions in average root mean squared error are 19%, 25%, 37%, and 29% over the benchmark methods. Then, it is used to predict yields of the considered crops in Australia towards 2025 under six different climate change scenarios. It is observed that although climate change has some boosting impact on crop yield, it is not sustainable to meet the demand. However, it is possible to keep crop yields rising while mitigating climate change.

Photosvaltable Characteristics Monitoring System Using the Internet of Things (IOT)


 
 
 
 Continuous monitoring of solar panel performance is essential to ensure effective and efficient generation of electrical energy. Developing a monitoring system such as an I-V characteristic curve tracker enables real-time monitoring and identification of potential problems or efficiency improvements. The system utilizes a circuit switching system for accurate and continuous measurement, providing insight into how the solar panel responds to changing environmental and load conditions. The data recorded by temperature, light intensity, current, and voltage sensors on the solar panel provides valuable information about its condition and the environmental factors affecting it. Utilizing IoT technology allows real-time monitoring of the data, accessible via a web page and presented graphically. The data is also stored in Excel format for further analysis, so users can access and monitor continuously as long as they are connected to the internet. From the experimental results, the Voc of the solar panel ranged from 18.7 V to 20.42 V, and Isc between 0.0727 A to 0.2145 A, which were tested under solar irradiation ranging from 100 to 500 W/m². This research demonstrates the importance of continuous monitoring and the effectiveness of the developed monitoring system in providing real-time data for analysis and optimization of solar panel performance. The system has been tested to meet the IEC No. 13B-23 standard for accuracy classes 0.2 and 0.5. A comparison between manual testing and sensor testing shows that despite the difference in accuracy levels, the sensor measurements have a relatively low average error (1.04% for voltage and 0.64% for current) compared to manual measurements, indicating that the sensor is reliable in providing accurate data. This demonstrates the reliability and effectiveness of using the sensor in the solar panel characteristics monitoring system.
 
 
 

New ultra-low cycle fatigue model for metal alloys

Ultra-low cycle fatigue failure is an important ultimate limit state of the structures. Based on the ductile fracture model proposed by the authors, the Coffin-Manson model, and the Miner law, a new model was proposed in this study. The strategy of the proposed model is intended for the prediction of ultra-low cycle fatigue fracture onset covering a wide range of stress triaxiality and being suitable for the variable amplitude straining. Three kinds of metal alloys, Q345qC steel, G20Mn5QT cast steel, and 2024-T351 aluminum alloy, were used to verify the model. The results showed that the proposed model advanced the original Coffin-Manson model greatly. More importantly, the verification demonstrates that the proposed model can predict the ultra-low cycle fatigue fracture initiation over a wide range of stress triaxiality and under variable amplitude straining. Besides, compared with the ULCF-2021 model, the proposed model has a relatively lower average error.

Intelligent Integrated System for Fruit Detection Using Multi-UAV Imaging and Deep Learning.

In the context of Industry 4.0, one of the most significant challenges is enhancing efficiency in sectors like agriculture by using intelligent sensors and advanced computing. Specifically, the task of fruit detection and counting in orchards represents a complex issue that is crucial for efficient orchard management and harvest preparation. Traditional techniques often fail to provide the timely and precise data necessary for these tasks. With the agricultural sector increasingly relying on technological advancements, the integration of innovative solutions is essential. This study presents a novel approach that combines artificial intelligence (AI), deep learning (DL), and unmanned aerial vehicles (UAVs). The proposed approach demonstrates superior real-time capabilities in fruit detection and counting, utilizing a combination of AI techniques and multi-UAV systems. The core innovation of this approach is its ability to simultaneously capture and synchronize video frames from multiple UAV cameras, converting them into a cohesive data structure and, ultimately, a continuous image. This integration is further enhanced by image quality optimization techniques, ensuring the high-resolution and accurate detection of targeted objects during UAV operations. Its effectiveness is proven by experiments, achieving a high mean average precision rate of 86.8% in fruit detection and counting, which surpasses existing technologies. Additionally, it maintains low average error rates, with a false positive rate at 14.7% and a false negative rate at 18.3%, even under challenging weather conditions like cloudiness. Overall, the practical implications of this multi-UAV imaging and DL-based approach are vast, particularly for real-time fruit recognition in orchards, marking a significant stride forward in the realm of digital agriculture that aligns with the objectives of Industry 4.0.

Contactless Respiration Monitoring Using Wi-Fi and Artificial Neural Network Detection Method.

Detecting respiration in a non-intrusive manner is beneficial not only for convenience but also for cases where the traditional ways cannot be applied. This paper presents a novel simple low-cost system where ambient Wi-Fi signals are acquired by a third-party tool (Nexmon) installed in a Raspberry Pi and is able to detect the respiration time domain waveform of a person. This tool was selected as it uses 80 MHz bandwidth of the Wi-Fi signal and supports the latest implementations that are widely used, such as 802.11ac. A neural network is developed to detect the respiration frequency of the waveform. Generated waves emulating respiration waveforms were used for training, validating, and testing the model. The model can be applied to unseen real measurement data and successfully determine the breathing frequency with a very low average error of 4.7% tested in 20 measurement datasets.

Carbon emissions of power transmission and transformation projects in the whole life cycle for smart sustainable energy systems

The study investigates the optimization of life cycle carbon emissions in smart sustainable energy systems through power transformation and transmission project power load predictions. Firstly, a multi-task learning-based short-term user load forecasting technique is developed, where the power load curves of multiple residential customers are grouped and classified using the K-means clustering method. Additionally, the Bidirectional Long Short-Term Memory (BiLSTM) technique is introduced to anticipate the power load intelligently. Secondly, a life cycle carbon emission assessment model for the power transmission and transformation project (PTTP) is constructed based on the life cycle assessment (LCA) method, which divides the project's life cycle into four stages: production, installation and construction, operation and maintenance, and demolition. Finally, an experimental evaluation of this model is conducted. The results demonstrate that compared with the baseline model Long Short-Term Memory (LSTM), this model achieves a significantly lower average Mean Absolute Error (MAE) at 3.62% while achieving significantly higher accuracy in power load forecasting at 94.34%. A comprehensive examination of carbon emissions across all four phases reveals that overall carbon emissions are highest during the operation and maintenance stage followed by the equipment production stage and installation/construction stage, with the lowest overall carbon emissions observed. Hence, this study endeavors to forecast power load demand with precision and identify the principal determinants of carbon emissions in power engineering. By discerning and managing these key factors, an optimal, energy-efficient intelligent power load scheme can be derived.

Multimodal audiovisual speech recognition architecture using a three‐feature multi‐fusion method for noise‐robust systems

AbstractExposure to varied noisy environments impairs the recognition performance of artificial intelligence‐based speech recognition technologies. Degraded‐performance services can be utilized as limited systems that assure good performance in certain environments, but impair the general quality of speech recognition services. This study introduces an audiovisual speech recognition (AVSR) model robust to various noise settings, mimicking human dialogue recognition elements. The model converts word embeddings and log‐Mel spectrograms into feature vectors for audio recognition. A dense spatial–temporal convolutional neural network model extracts features from log‐Mel spectrograms, transformed for visual‐based recognition. This approach exhibits improved aural and visual recognition capabilities. We assess the signal‐to‐noise ratio in nine synthesized noise environments, with the proposed model exhibiting lower average error rates. The error rate for the AVSR model using a three‐feature multi‐fusion method is 1.711%, compared to the general 3.939% rate. This model is applicable in noise‐affected environments owing to its enhanced stability and recognition rate.

Affective Modelling of Eustress and Distress using Psychological Scales

This article is a case study that illustrates how a linear regression model can be implemented in eustress and distress analysis based on the correlation between emotion and stress and uses it to develop prediction equations of stress. This study proposes the use of five questionnaires; Perceived Stress Scale 10, Academic Eustress scale, Academic Distress scale, Bosse’s Distress Eustress scale and Adolescent Distress Eustress scale to determine perceived stress, eustress or distress. A sixth questionnaire, the Self-Assessment Manikin was used to determine emotional state in terms of valence and arousal, which are represented on 2-dimensional axis, where the x axis represents valence, and the y axis represents arousal. An analysis of the relationship between the results of the stress questionnaires and results of SAM based valence and arousal is carried out. Significant correlations are then used to derive regression equations used to predict eustress, distress or perceived stress. The findings showed that neither valence nor arousal was correlated with perceived stress, hence no regression equation was derived for it. However, valence and/or arousal were correlated with the remaining five questionnaires. Finally, this article analyzes the predictions comparing actual vs predicted values. Error analysis showed that the ADES questionnaire had the lowest average error, making it the most suitable in predicting eustress and distress from emotion.

Modeling the grain yield loss and quality assessment of some durum wheat (Triticum durum Desf.) genotypes under semi-arid conditions

This study was conducted at the Experimental Station for Field Crops ITGC in Setif -Algeria during the growing season (2020-2021), to evaluate the parameters of the technological quality efficiency of three durum wheat genotypes {Boussellam (BOS), Oued El Bared (OB), and GTA dur (GTA)} and the efficiency of using mathematical models based on the relation between grain moisture in the field and grain yield to estimate grain yield loss caused by the delayed harvest according to the randomized blocks design with three replications. Results demonstrated that the genotype effects were significant for all technological parameters, the protein content varied from 13.70 to 15.4%; the highest content of protein registered by (OB) was 15.4%. In addition, the values of test weight varied between 79.47 and 81.97 kg hL-1, with a general mean of 80.96 kg hL-1. The study of correlations test showed that there was a significant and positive correlation between the decreased grain moisture in the field and the loss in final grain yield, which suggests that the final grain yield decreases as the grain moisture decreases. This loss can be predicted using a mathematical regression model. The statistical analysis revealed the best agreement between measured and simulated grain yield, with low average absolute error and root mean square error. The grain yield was also well simulated with the observed yield giving a coefficient of efficiency (E) of 0.76, i.e., with a simulation capacity of 76%. Overall, and after the physiological maturity of the grains the mathematical model proved that with the 1% loss of grain moisture, there is a loss of about 0.290 t ha-1 of grain yield.