Accuracy Of Artificial Neural Network Research Articles

Rock mass quality classification based on deep learning: A feasibility study for stacked autoencoders

Objective and accurate evaluation of rock mass quality classification is the prerequisite for reliable stability assessment. To develop a tool that can deliver quick and accurate evaluation of rock mass quality, a deep learning approach is developed, which uses stacked autoencoders (SAEs) with several autoencoders and a softmax net layer. Ten rock parameters of rock mass rating (RMR) system are calibrated in this model. The model is trained using 75% of the total database for training sample data. The SAEs trained model achieves a nearly 100% prediction accuracy. For comparison, other different models are also trained with the same dataset, using artificial neural network (ANN) and radial basis function (RBF). The results show that the SAEs classify all test samples correctly while the rating accuracies of ANN and RBF are 97.5% and 98.7%, repectively, which are calculated from the confusion matrix. Moreover, this model is further employed to predict the slope risk level of an abandoned quarry. The proposed approach using SAEs, or deep learning in general, is more objective and more accurate and requires less human intervention. The findings presented here shall shed light for engineers/researchers interested in analyzing rock mass classification criteria or performing field investigation.

Modeling pedestrian behavior at urban signalised intersections using statistical-ANN hybrid approach – Case study of New Delhi

Traffic crashes at intersections are major hazard all over the world. The victims of road crashes are mostly pedestrians. A large share of accidents can be related to unsafe crossing behavior by pedestrians. The accidents are mostly fatal for pedestrians especially if the unsafe act is signal violation. This study primarily focused on cognizing factors which are significantly affecting the signal violation behavior of pedestrians’ at urban signalised intersections. Data is collected at 11 signalised intersections in New Delhi, India using video recording technique. Significant variables are identified using conventional statistics. The variables are then used as input neuron for the Artificial Neural Network (ANN) for modeling signal violation behavior. Sensitivity analysis is performed to identify the relative importance of factors related to the pedestrian signal violation. Host of different factors such as pedestrian demographic, behavioral attributes, crossing state and mobile usage are notably influencing signal violations. Pedestrian crossing speed, crossing path and waiting time are the top three predictors of violation behavior. The accuracy of ANN to predict signal violation behaviour is found to be about 85% and is considerably higher than conventional binary logistic regression (BLR) model. The area under the receiver operating characteristic (ROC) curve (AUC) is found to be 0.753, suggesting good model performance. The results suggest that ANN is a robust alternative to conventional regression, for studying and modeling pedestrian behavior. The findings from this study would assist engineers and policy makers to take proactive measures for designing pedestrian friendly facilities to reduce signal violations. This would ultimately improve pedestrian safety at urban signalised intersections.

Towards a generalised artificial neural network for sub-grid filtered density function closure in turbulent combustion

An artificial neural network (ANN) is trained on moderate or intense low-oxygen dilution (MILD) combustion to predict the sub-grid filtered density function (FDF) in large eddy simulation (LES). For wide usability, a new quantity ψ is calculated by scaling the filtered mixture fraction using the stoichiometric value and a logarithmic function, which is used as an input feature of the ANN. Self-testing on MILD combustion is first conducted for validation and determining the best layout of ANN. An ANN with 4 hidden layers and the activation function of rectified linear unit (ReLU) has the highest accuracy. This ANN is subsequently tested on five different premixed combustion cases. Overall predictions of the progress variable FDF are satisfactory, and the filtered chemical source term modelled by the FDFs is in good agreement with the DNS data for each case. However, an underprediction of the filtered reaction rate occurs for the hydrogen–air flames, and the ANN accuracy is lower than that of a presumed β-FDF approach. DNS data of one premixed flame are then added to the training datasets, and newly trained ANNs are tested on remaining premixed flame cases. When the testing cases are similar to the added training dataset, improved predictions of marginal FDFs are observed, and the modelled progress variable source terms show good comparisons with the corresponding DNS data. Compared to the presumed β-FDF method, the new ANN shows an improved accuracy and serves as a viable alternative for FDF-based combustion modelling in the LES of turbulent flames.

Caenorhabditis elegans Connectomes of both Sexes as Image Classifiers

Connectome, the complete wiring diagram of the nervous system of an organism, is the biological substrate of the mind. While biological neural networks are crucial to the understanding of neural computation mechanisms, recent artificial neural networks (ANNs) have been developed independently from the study of real neural networks. Computational scientists are searching for various ANN architectures to improve machine learning since the architectures are associated with the accuracy of ANNs. A recent study used the hermaphrodite Caenorhabditis elegans (C. elegans) connectome for image classification tasks, where the edge directions were changed to construct a directed acyclic graph (DAG). In this study, we used the whole-animal connectomes of C. elegans hermaphrodite and male to construct a DAG that preserves the chief information flow in the connectomes and trained them for image classification of MNIST and fashion-MNIST datasets. The connectome-inspired neural networks exhibited over 99.5% and 92.6% of accuracy for MNIST and fashion-MNIST datasets, respectively, which increased from the previous study. Together, we conclude that realistic biological neural networks provide the basis of a plausible ANN architecture. This study suggests that biological networks can provide new inspiration to improve artificial intelligences (AIs).

Differential Diagnosis Models for Multiple Myeloma Combined with Renal Injury and Chronic Kidney Disease or Nephrotic Syndrome

Background: When multiple myeloma(MM) is combined with renal injury, most patients are easily misdiagnosed as kidney diseases. This study aimed to establish a differential diagnosis model for MM combined with renal injury based on clinical information. Methods: A total of 77 patients with MM combined with renal injury were recruited as the case group, and 112 patients with kidney diseases were recruited as a control group. Support vector machine (SVM), decision tree (DT), and artificial neural network (ANN) models were developed based on significant clinical variables. Accuracy and area under the receiver operating characteristic curve (AUC) were used to evaluate each model. Results: Accuracies of SVM, DT, and ANN were 0.843,0.902, and 0.941. The AUCs of SVM, DT, and ANN were 0.822,0.879, and 0.932. Lower extremity edema, bone pain, and lactate dehydrogenase (LDH) were common important indicators identified by SVM, DT and ANN models. When these three indicators were excluded, the ANN model prediction effect decreased significantly (P<0.05). Conclusion: The results suggest that the ANN model best predicts the differential diagnosis between MM combined with renal injury and chronic kidney disease/nephrotic syndrome. Important features contributing to identifying the diseases, including lower extremity edema, bone pain, and LDH, may assist in diagnosing such diseases in the future.

Non-destructive method for identification and classification of varieties and quality of coffee beans based on soft computing models using VIS/NIR spectroscopy

Coffee is one of the most popular and frequently consumed beverages on the planet. Coffee has a significant commercial value, estimated to be in the billions of dollars and consumption has risen steadily over the last two decades. Near-infrared spectroscopy is one of the non-destructive optical technologies for the evaluation of agricultural products to identify food adulteration. Thus, it is an interesting and worthwhile subject to research and study. In this research, a near-infrared spectroscopy approach along with statistical methods of principal component analysis (PCA), partial-least-squares regression (PLSR), latent dirichlet allocation (LDA), and artificial neural network (ANN) as a fast and non-destructive method was used with to detect and classify coffee beans using reference data obtained by gas chromatography–mass spectrometry (GC–MS). Results showed that the accuracy of PLSR, LDA, and ANN while our reference data was palmitic acid, respectively were 97.3%, 97.92%, and 97.3% and while reference data was caffeine, accuracy results were 94.71%, 95.83%, and 98.96%, respectively.

To Establish an Early Prediction Model for Acute Respiratory Distress Syndrome in Severe Acute Pancreatitis Using Machine Learning Algorithm.

To develop binary and quaternary classification prediction models in patients with severe acute pancreatitis (SAP) using machine learning methods, so that doctors can evaluate the risk of patients with acute respiratory distress syndrome (ARDS) and severe ARDS at an early stage. A retrospective study was conducted on SAP patients hospitalized in our hospital from August 2017 to August 2022. Logical Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), and eXtreme Gradient Boosting (XGB) were used to build the binary classification prediction model of ARDS. Shapley Additive explanations (SHAP) values were used to interpret the machine learning model, and the model was optimized according to the interpretability results of SHAP values. Combined with the optimized characteristic variables, four-class classification models, including RF, SVM, DT, XGB, and Artificial Neural Network (ANN), were constructed to predict mild, moderate, and severe ARDS, and the prediction effects of each model were compared. The XGB model showed the best effect (AUC = 0.84) in the prediction of binary classification (ARDS or non-ARDS). According to SHAP values, the prediction model of ARDS severity was constructed with four characteristic variables (PaO2/FiO2, APACHE II, SOFA, AMY). Among them, the overall prediction accuracy of ANN is 86%, which is the best. Machine learning has a good effect in predicting the occurrence and severity of ARDS in SAP patients. It can also provide a valuable tool for doctors to make clinical decisions.

Analysis And Comparison Of Ventricular Cardiac Arrhythmia Classification Using Potassium Channel Parameters With ANN And KNN Classifier

Aim: The Motive of this research is to analyze, compare ventricular Cardiac Arrhythmia (CA) classification using potassium channel (k+) parameters with Artificial Neural Network (ANN) and K-Nearest Neighbor (KNN) classifiers. Materials and Methods: D Noble Model For Human Ventricular Tissue (DNFHVT) is used for our classification. The DNFHVT is a mathematical model of action potential focusing on major ionic currents like K+,Na+ and Ca+.. Size of the sample was calculated by keeping threshold 0.05, G Power 80%, confidence interval 95% and enrolment ratio as 1. Number of samples considered is 20. These data are imported to KNN and ANN classifiers to find better accuracy among them. The accuracy of novel ANN and KNN classifiers for 20 samples is obtained by alternating the cross fold validation. These results will be imported to Statistical Package for the Social Science (SPSS) software to identify the overall accuracy for each classifier. Results: The results are obtained from SPSS for novel ANN and KNN classifiers. ANN shows accuracy of 13.14% with standard deviation (1.6800) and Standard error mean (0.3757). Similarly KNN produces an accuracy value of 7.19% with standard deviation (1.6902) and Standard error mean (0.377). Conclusion: As of the results, it clearly shows that ANN has better accuracy for classification than KNN.

Use of artificial neural networks in the prognosis of musculoskeletal diseases—a scoping review

To determine the current evidence on artificial neural network (ANN) in prognostic studies of musculoskeletal diseases (MSD) and to assess the accuracy of ANN in predicting the prognosis of patients with MSD. The scoping review was reported under the Preferred Items for Systematic Reviews and the Meta-Analyses extension for Scope Reviews (PRISMA-ScR). Cochrane Library, Embase, Pubmed, and Web of science core collection were searched from inception to January 2023. Studies were eligible if they used ANN to make predictions about MSD prognosis. Variables, model prediction accuracy, and disease type used in the ANN model were extracted and charted, then presented as a table along with narrative synthesis. Eighteen Studies were included in this scoping review, with 16 different types of musculoskeletal diseases. The accuracy of the ANN model predictions ranged from 0.542 to 0.947. ANN models were more accurate compared to traditional logistic regression models. This scoping review suggests that ANN can predict the prognosis of musculoskeletal diseases, which has the potential to be applied to different types of MSD.

An Uncertainty Analysis on Finite Difference Time-Domain Computations With Artificial Neural Networks: Improving accuracy while maintaining low computational costs

Artificial neural networks (ANNs) have appeared as a potential alternative for uncertainty quantification (UQ) in the finite difference time-domain (FDTD) computation. They are applied to build a surrogate model for the computation-intensive FDTD simulation and to bypass the numerous simulations required for UQ. However, when the surrogate model utilizes an ANN, a considerable number of data are generally required for high accuracy, and generating such large quantities of data becomes computationally prohibitive. To address this drawback, a number of adaptations for ANNs are proposed, which additionally improves the accuracy of ANNs in UQ for the FDTD computation while maintaining a low computational cost. The proposed algorithm is tested for application in bioelectromagnetics, and considerable speed up, as well as the improved accuracy of UQ, is observed compared to traditional methods such as the nonintrusive polynomial chaos (NIPC) method.

Extremum hybrid intelligent-inspired models for accurate predicting mechanical performances of turbine blisk

In complex aeroengine structures, it is necessary to consider multi-physical loads involving vibrational, structural, aerodynamical and heat loads for safe design. Turbine blisk commonly involves multi-failure modes such as stress, strain, deformation, fatigue, and creep under multi-physical loads during operation, so that efficient and accurate simulating approach is one of main issues in structural safety design. Machine learning (ML) approaches using artificial intelligence (AI) are workable to improve the simulation efficiency of mechanical responses under nonlinear dynamic structural analysis. In this work, hybrid artificial neural network (ANN) models are proposed to simulate the failure modes of turbine blisk. The accuracy of AI-based ANN is discussed using six music-inspired optimization algorithms named as harmony search (HS), improved HS (IHS), global-best HS (GHS), improved GHS (IGHS), adaptive GHS (AGHS) and Gaussian GHS (GGHS). Herein, these six music-inspired optimization algorithms are employed to find the undetermined coefficients and hyperparameters for the ANN models, and two adjusting strategy is introduced to explore the optimal values of the undetermined coefficients and hyperparameters in the IGHS, AGHS and GGHS. These hybrid models are used to perform the approximation of dimensional mechanical capacities such as stress, strain and deformation of aeroengine structures. These hybrid models are compared by using tendency, accuracy and agreement for failure modes of turbine system. It was conducted that the music-inspired basis harmony optimization procedures can provide the acceptable nonlinear connection between input and output responses of aeroengine turbine blisk in training phase of ANN models. The GGHS algorithm shows superior performance with the highest accuracy and tendency among other HS optimization algorithms.

Predicting Cu(II) Adsorption from Aqueous Solutions onto Nano Zero-Valent Aluminum (nZVAl) by Machine Learning and Artificial Intelligence Techniques

Predicting the heavy metals adsorption performance from contaminated water is a major environment-associated topic, demanding information on different machine learning and artificial intelligence techniques. In this research, nano zero-valent aluminum (nZVAl) was tested to eliminate Cu(II) ions from aqueous solutions, modeling and predicting the Cu(II) removal efficiency (R%) using the adsorption factors. The prepared nZVAl was characterized for elemental composition and surface morphology and texture. It was depicted that, at an initial Cu(II) level (Co) 50 mg/L, nZVAl dose 1.0 g/L, pH 5, mixing speed 150 rpm, and 30 °C, the R% was 53.2 ± 2.4% within 10 min. The adsorption data were well defined by the Langmuir isotherm model (R2: 0.925) and pseudo-second-order (PSO) kinetic model (R2: 0.9957). The best modeling technique used to predict R% was artificial neural network (ANN), followed by support vector regression (SVR) and linear regression (LR). The high accuracy of ANN, with MSE < 10−5, suggested its applicability to maximize the nZVAl performance for removing Cu(II) from contaminated water at large scale and under different operational conditions.

Retrieving leaf area index of rubber plantation in Hainan Island using empirical and neural network models with Landsat images

The leaf area index (LAI) is an important parameter for describing the growth status and canopy structure of vegetation. The rapid and accurate acquisition of the vegetation or agroforestry LAI has great scientific significance in agroforestry ecological ecosystems research and a very important practical value for guiding agricultural and forestry production. In this study, the typical tropical crops (rubber forest) in Hainan Island were selected as the research area, the empirical and neural network (NN) LAI estimation models of rubber forest were constructed based on satellite remote sensing vegetation indices and the field LAI measurement data, and the spatiotemporal variation was analyzed. The results showed that, compared with normalized difference vegetation index (NDVI), green NDVI (GNDVI), ratio VI (RVI), normalized near-infrared (NNIR), wide dynamic range VI (WDRVI), and normalized difference water index (NDWI), enhanced vegetation index (EVI), soil adjusted vegetation index (SAVI), DVI, renormalized DVI (RDVI), and modified SAVI (MSAVI) have higher correlations with LAI. Among the LAI estimation models of rubber forest based on empirical and artificial NN (ANN) models, the estimation accuracy of ANN achieves the highest value. The linear fitting determination coefficient R2 of the observed and simulated rubber forest LAI was 0.85 (p < 0.001), the root mean square error (RMSE) was 0.15, and the average relative error (RE) was 1.93%. However, there was underestimation in the middle-value area and overestimation in the high- and low-value areas of LAI. Based on remote sensing mapping of the rubber forest LAI, the high LAI values (4.40 to 6.00 m2 m − 2) were mainly distributed in Danzhou and Baisha (west of Hainan Island); the middle LAI values (3.80 to 4.40 m2 m − 2) were mainly located in Chengmai, Tunchang, and Qiongzhong (middle of Hainan Island); and the low LAI values (<3.80 m2 m − 2) were shown primarily on Ding’an, Qionghai, Wanning, Ledong, and Sanya (east and south of Hainan Island). In summary, the remote sensing estimation model for the rubber plantation LAI based on the vegetation index has high accuracy and good values for application.

НЕЙРОСЕТЕВОЙ АНАЛИЗ ЦИФРОВЫХ ИЗОБРАЖЕНИЙ ЛЕЙКОЦИТОВ ДЛЯ ОБНАРУЖЕНИЯ ЗАБОЛЕВАНИЙ КРОВИ

The article discusses the application of the transfer learning method for the ensemble of artificial convolutional neural networks with preliminary digital image segmentation for blood cells in order to classify them later. The results obtained during neural networks classification demonstrate the efficiency of such technologies used for improving the accuracy of artificial neural networks when solving the problems of medical images segmentation for leukocytes in order to diagnose hematologic diseases.

Developing the Security for Cloud Information Via Alexnet Learning Model versus the accuracy of Artificial Neural Network

The main objective of the study is to protect cloud from different types of attacks by using AlexNet classifier compared accuracy with Artificial Neural Network and user’s data to be stored in the cloud safely by Advanced Encryption Standard. Materials and Methods: This research examines two groups AlexNet withArtificial Neural Network. Statistical study used 1300 training and 403 testing datasets from UNSW-NB15 dataset. ClinCalc programme utilised N=10, 0.05 is alpha value, 0.8% is G-Power, and 95% confidence interval.Result and Discussion: Novel Alexnet (91.081%) has an increased precision over ANN (90.075%) with the P value is 0.012(p&lt;0.05) from the results of Independent samples t-test. There is a statistical significant difference between these two algorithms.Conclusion: This study concludes that the Novel AlexNet classifier algorithm seems fundamentally better than the ANN in terms of increasing the accuracy of secure cloud data and holding sensitive data with the dataset of UNSW-NB15.

The accuracy of artificial neural networks and logit models in predicting the companies&amp;#8217; financial distress

The accuracy of artificial neural networks and logit models in predicting the companies&amp;#8217; financial distress

Enhancing the Classification Accuracy of Rice Varieties by Using Convolutional Neural Networks

The aim of this study is to enhance the classification accuracy of rice varieties that are quite similar in external observation. In this study, 17 rice grain varieties popularly planted in Vietnam are classified with an Artificial Neural Network (ANN) and Convolutional Neural Network (CNN) models. The two CNN models (modified VGG16 and modified ResNet50) are based on pre-trained VGG16 and Resnet50 models. Two datasets are used in the experiments: a feature dataset extracted using an extended improved local ternary pattern (extended ILTP) method, and an image dataset generated with a data augmentation technique. The feature dataset was fed into the ANN, while the image dataset was fed into the CNN models. The highest classification accuracies of ANN, modified VGG16, and modified ResNet50 models are 92.82%, 96.41%, and 97.88%, respectively. The results show that the modified VGG16 and ResNet50 models significantly improved classification accuracy of the 17 varieties of rice. In addition, the experiments show that the dimensions of the image dataset can affect the performance of the CNN models. This research can be developed for applications of rice varieties classification and identification.

Forecasting of Canopy Temperatures Using Machine Learning Algorithms

Highlights This study analyzes the feasibility of using Artificial Neural Networks (ANNs) to estimate canopy temperatures. A methodology is introduced to forecast canopy temperatures using historical canopy temperatures. ANNs can predict canopy temperatures with satisfactory accuracy for plant stress-based irrigation scheduling. The methodology can be useful to add redundancy to plant stress-based irrigation scheduling methods. Abstract. Recent advances can provide farmers with irrigation scheduling tools based on crop stress indicators to assist in the management of Variable Rate Irrigation (VRI) center pivot systems. These tools were integrated into an Irrigation Scheduling Supervisory Control and Data Acquisition System (ISSCADAS) developed by scientists with the USDA-Agricultural Research Service (ARS). The ISSCADAS automates the collection of data from a network of wireless infrared thermometers (IRTs) distributed on a center pivot’s lateral and in the field irrigated by the center pivot, as well as data from a wireless soil water sensor network and a microclimate weather station. This study analyzes the use of Artificial Neural Networks (ANNs), a type of machine learning algorithm, for the forecasting of canopy temperatures obtained by a wireless network of IRTs mounted on a three-span VRI center pivot irrigating corn near Bushland, TX, during the summer of 2017. Among the predictors used by the ANNs were weather variables relevant to the estimation of evapotranspiration (i.e., air temperature, relative humidity, solar irradiance, and wind speed), irrigation management variables (irrigation treatment, irrigation scheduling method, and the amount of water received during the last 5 days as irrigation or rainfall), and days after planting. Two case studies were conducted using data collected from periodic scans of the field performed during the growing season by running the pivot dry. In the first case, data from the first three scans were used to train an ANN, and canopy temperatures estimated using the ANN were then compared against canopy temperatures measured by the network of IRTs during the fourth scan. In the second case, data from the first six scans were used to train ANNs, and canopy temperatures estimated using the ANN were then compared against canopy temperatures measured by the network of IRTs during the seventh scan. The Root of the Mean Squared Error (RMSE) of ANN predictions in the first case ranged from 1.04°C to 2.49°C, whereas the RMSE of ANN predictions in the second case ranged from 2.14°C to 2.77°C. To assess the impact of ANN accuracy on irrigation management, estimated canopy temperatures were fed to a plant-stress-based irrigation scheduling method, and the resulting prescription maps were compared against prescription maps obtained by the same method using the canopy temperatures measured by the network of IRTs. In the first case, no difference was found between both prescription maps. In the second case, only one plot (out of 26) was assigned a different prescription. Results of this study suggest that machine learning techniques can be used to assist the ISSCADAS in situations where canopy temperatures cannot be measured by the network of IRTs due to poor visibility conditions, or because the center pivot cannot traverse the field within a reasonable amount of time. Keywords: Artificial Neural Network, Canopy temperature sensing, Center pivot irrigation, Irrigation scheduling, Machine learning, Metamodeling, Variable rate irrigation.

A movie box office revenues prediction algorithm based on human-machine collaboration feature processing

Improving the accuracy of box office revenue forecasts is conducive to stimulating the creation, market investment, infrastructure construction, and rational allocation of public resources in the film market, as well as promoting social welfare and cultural prosperity. Since the existing box office revenue prediction algorithm does not consider film industry structure, the prediction accuracy is not satisfying. This paper firstly builds a two-stage human-machine collaborative feature processing framework. In the first stage, based on the box office data, the regression decision tree algorithm is used to process all the box office features preliminarily and delete the unimportant features automatically. In the second stage, feature processing is coupled with the built Artificial Neural Network (ANN). In this stage, the features processed by the machine are manually classified, and multiple, incompatible feature sets are divided. After designing the incompatible set network pruning algorithm, the neural network is pruned. We construct the data set with a total of 7098 movies crawled online on four platforms. Numerical experimental results show that the Mean Absolute Error (MAE) of the two-stage algorithm is significantly better than the baseline model, which can effectively reduce the noise caused by encoding between incompatible features directly, improve the prediction accuracy of ANN, accelerate the forward inference speed of ANN and reduce the consumption of computing resources.

Optimizing model parameters of artificial neural networks to predict vehicle emissions

This paper presents a novel approach to predict carbon dioxide (CO2), nitrogen oxides (NOx), and carbon monoxide (CO) emissions of diesel vehicles using artificial neural network (ANN), which offer high degrees of accuracy and practicality. Six operating parameters (velocity, engine speed, engine torque, engine coolant temperature, fuel/air ratio, and intake air mass flow) collected through on-board diagnostic interface were used as predictors of exhaust emissions. The importance of each parameter to the emission predictions were comprehensively analyzed by comparing the coefficient of determination, root mean square error, cumulative emissions, and instantaneous emission rates. The emission prediction accuracy of ANN tends to increase as more parameters were considered as model inputs at the same time. However, the level of accuracy improvement depends on the input parameters. For CO2 emissions, engine torque and fuel/air ratio were good predictors for achieving high prediction accuracy. The relative importance of intake air mass flow rate and fuel/air ratio was high for NOx and CO predictions, respectively. In addition, the emission prediction accuracy of ANN depends on the vehicle type (Euro 5, Euro 6b, Euro 6d-temp). The emission prediction accuracy of vehicles equipped with after-treatment devices (selective catalytic reduction and lean NOx trap) was lower than that of vehicles without after-treatment devices.