Function Neural Network Research Articles

A novel construction and evaluation framework for driving cycle of electric vehicles based on energy consumption and emission analysis

The driving cycle (DC) is essential for establishing vehicle emission standards, transportation policies, and urban planning. However, existing driving cycles demonstrate poor representativeness and excessive randomness due to the insufficient capture of driving characteristics. Therefore, a novel framework for constructing and evaluating driving cycles of electric vehicles (EVs) based on energy consumption and emissions analysis is proposed to enhance the representativeness of the constructed driving cycles. First, based on road information, an improved dual-chain Markov chain method combined with the self-organizing mapping (SOM) neural network is introduced for clustering and constructing driving cycles. Subsequently, a double-layer evaluation model oriented towards energy consumption and emissions is established through a combination of model-driven and data-driven approaches to select a representative driving cycle (RDC). Finally, comparative experiments are conducted to evaluate the feasibility and scientific validity of the proposed method in multiple dimensions. The results indicate that the driving cycle constructed in this study demonstrates excellent representativeness, with an emission error of 2.04% and a comprehensive characterization parameter (CCP) value of 0.097. This study emphasizes the necessity of incorporating reasonable constraints in the driving cycle construction. This improved representativeness provides a reliable foundation for electric vehicle design and transportation policy development.

From One-dimensional to Multidimensional Map Neural Networks

In the contemporary landscape, artificial neural networks transcend their traditional role of function approximation and have found diverse applications in fields such as image classification, machine translation, speech recognition, and natural language processing. In some datasets, traditional architectures exhibit low test and training accuracy, coupled with high loss and prolonged training times. This study aims to introduce innovative neural network architectures that outperform conventional models. The research presents a novel framework integrating one-dimensional and multidimensional map neural networks, consisting of three distinct architectures: 1D-Map, 2D-Map, and 3D-Map. A systematic performance comparison with traditional models is conducted following the implementation of these architectures. The evaluation spans four datasets, encompassing the domains of heat treatment of electroless Ni-P nano coatings, letter recognition, combined cycle power plants, and Seoul bike sharing demand. In the heat treatment dataset, the proposed 3D-Map architecture reached 0.055 higher average test accuracy than traditional MLP architecture. In the letter recognition dataset, the test accuracy of 3D-Map architecture was 0.0523 higher than the test accuracy of the LSTM architecture. In the combined cycle power plant dataset, the test accuracy of the 3D-Map architecture was 0.0612 more than the test accuracy of the MLP architecture. In the Seoul bike-sharing demand dataset, the test accuracy of the 2D-Map architecture was 0.0696 higher than the test accuracy of the LSTM architecture. The study's findings underscore the consistently superior performance of the proposed architectures compared to their traditional counterparts.

Boosting the Development and Management of Wind Energy: Self-Organizing Map Neural Networks for Clustering Wind Power Outputs

Aimed at the information loss problem of using discrete indicators in wind power output characteristics analysis, a self-organizing map neural network-based clustering method is proposed in this study. By identifying the appropriate representativeness and topological structure of the competition layer, cluster analysis of the wind power output process in four seasons is realized. The output characteristics are evaluated through multiple evaluation indicators. Taking the wind power output of the Hunan power grid as a case study, the results underscore that the 1 × 3-dimensional competition layer structure had the highest representativeness (72.9%), and the wind power output processes of each season were divided into three categories, with a robust and stable topology structure. Summer and winter were the most representative seasons. Summer had strong volatility and small wind power outputs, which required the utilization of other power sources to balance power supply and load demand. Winter featured low volatility and large wind power outputs, necessitating cooperation with peak-shaving power sources to enhance the power grid’s absorbability to wind power. The seasonal clustering analysis of wind power outputs will be helpful to analyze the seasonality of wind power outputs and can provide scientific and technical support for guiding the power grid’s operation and management.

A Method for Predicting the Impact of Labor Education at Institutions of Higher Education Based on a Two-Stage Clustering Algorithm

In order to develop more accurate teaching plans and methods based on students’ characteristics and needs, and improve the effectiveness of labor education, a two-stage clustering algorithm based method for predicting the effectiveness of labor education in universities is studied. A two-stage clustering algorithm-based method for predicting the effectiveness of labor education in universities has been proposed. Through in-depth analysis of labor education data in universities, principal component analysis (PCA) was used to select 16 most representative variables from the original features. Subsequently, self-organizing map (SOM) neural network was used for preliminary clustering, and the optimal number of clusters was determined to be 8. Furthermore, the K-means clustering algorithm performs fine clustering based on the initial centers provided by SOM, significantly improving the stability and accuracy of clustering. At the same time, processing the outliers in the clustering results has reduced their impact on the clustering effect. Finally, the processed data was input into an improved XGBoost prediction model, which achieved a true positive rate (TPR) of over 85% on the ROC curve when predicting the effectiveness of labor education in universities. While maintaining a low false positive rate (FPR), the model outperformed other comparison methods significantly. This achievement not only validates the effectiveness of the proposed method in feature selection, clustering optimization, and prediction model construction, but also confirms its ability to accurately predict the labor education effectiveness of students in different grades, providing strong support for higher education institutions to optimize teaching plans and improve educational outcomes.

Deep self-organizing map neural networks improve the segmentation for inadequate plantar pressure imaging data set

ABSTRACT This study introduces a deep self-organizing map neural network based on level-set (LS-SOM) for the customization of a shoe-last defined from plantar pressure imaging data. To alleviate the over-segmentation problem of images, which refers to segmenting images into more subcomponents, a domain-based segmentation model of plantar pressure images was constructed. The domain growth algorithm was subsequently modified by optimizing its parameters. A SOM with 10, 15, 20, and 30 hidden layers was compared and validated according to domain growth characteristics by using merging and splitting algorithms. Furthermore, we incorporated a level set segmentation method into the plantar pressure image algorithm to enhance its efficiency. Compared to the literature, this proposed method has significantly improved pixel accuracy, average cross-combination ratio, frequency-weighted cross-combination ratio, and boundary F1 index comparison. Using the proposed methods, shoe lasts can be designed optimally, and wearing comfort is enhanced, particularly for people with high blood pressure.

Research on typical operating conditions of hydrogen production system with off-grid wind power considering the characteristics of proton exchange membrane electrolysis cell

Hydrogen energy, with its abundant reserves, green and low-carbon characteristic, high energy density, diverse sources, and wide applications, is gradually becoming an important carrier in the global energy transformation and development. In this paper, the off-grid wind power hydrogen production system is considered as the research object, and the operating characteristics of a proton exchange membrane (PEM) electrolysis cell, including underload, overload, variable load, and start- stop are analyzed. On this basis, the characteristic extraction of wind power output data after noise reduction is carried out, and then the self-organizing mapping neural network algorithm is used for clustering to extract typical wind power output scenarios and perform weight distribution based on the statistical probability. The trend and fluctuation components are superimposed to generate the typical operating conditions of an off-grid PEM electrolytic hydrogen production system. The historical output data of an actual wind farm are used for the case study, and the results confirm the feasibility of the method proposed in this study for obtaining the typical conditions of off-grid wind power hydrogen production. The results provide a basis for studying the dynamic operation characteristics of PEM electrolytic hydrogen production systems, and the performance degradation mechanism of PEM electrolysis cells under fluctuating inputs.

Artificial intelligence in prostate cancer: The potential of machine learning models and neural networks to predict biochemical recurrence after robot-assisted radical prostatectomy

ABSTRACT Introduction: This study aimed to evaluate the usefulness of machine learning (ML) and neural network (NN) models versus traditional statistical methods for estimating biochemical recurrence (BCR) in men following robot-assisted radical prostatectomy (RARP). Methods: Patients who underwent RARP from November 2011 to July 2022 were taken in the study. Patients with BCR were assigned to Group 2, whereas those without BCR were placed in Group 1. Preoperative and postoperative parameters, together with demographic data, were recorded in the database. This study used one NN, the radial basis function NN (RBFNN), and two ML approaches, the K-nearest neighbor and XGboost ML models, to predict BCR. Results: Following the application of exclusion criteria, 516 patients were deemed eligible for the study. Of those, 234 (45.3%) developed BCR, and 282 (54.7%) did not. The results showed that the median follow-up period was 24 (15–42) months, and the median BCR diagnosis was 12.23 ± 15.58 months. The area under the curve (AUC) for the Cox proportional hazard analysis was 0.77. The receiver-operating characteristic curves (AUCs) for the XGBoost and K closest neighbor models were 0.82 and 0.69, respectively. The RBFNN’s AUC was 0.82. Conclusions: The classical statistical model was outperformed by XGBoost and RBFNN models in predicting BCR.

Multi-AUV Kinematic Task Assignment Based on Self-Organizing Map Neural Network and Dubins Path Generator.

To deal with the task assignment problem of multi-AUV systems under kinematic constraints, which means steering capability constraints for underactuated AUVs or other vehicles likely, an improved task assignment algorithm is proposed combining the Dubins Path algorithm with improved SOM neural network algorithm. At first, the aimed tasks are assigned to the AUVs by the improved SOM neural network method based on workload balance and neighborhood function. When there exists kinematic constraints or obstacles which may cause failure of trajectory planning, task re-assignment will be implemented by changing the weights of SOM neurals, until the AUVs can have paths to reach all the targets. Then, the Dubins paths are generated in several limited cases. The AUV's yaw angle is limited, which results in new assignments to the targets. Computation flow is designed so that the algorithm in MATLAB and Python can realize the path planning to multiple targets. Finally, simulation results prove that the proposed algorithm can effectively accomplish the task assignment task for a multi-AUV system.

Adaptive formation control for obstacle avoidance of USVs with asymmetric input saturation

Due to the complex maritime navigation environment, Unmanned Surface Vessels (USVs) are influenced by unknown nonlinear dynamics arising from external disturbances and internal uncertainties. Achieving effective formation control while maintaining obstacle avoidance performance presents significant challenges. This article proposes a Neural Networks (NNs) adaptive formation Artificial Potential Field (APF) obstacle avoidance control method for multiple USVs. By employing online updates of Radial Basis Function (RBF) NNs technology, the unknown nonlinear dynamics are approximated, thus addressing complex nonlinear dynamics problems. In scenarios involving multiple USVs navigating under high wind and wave conditions, collisions with obstacles frequently occur. To tackle this issue, a leader-follower control strategy is designed that effectively addresses risk assessment and obstacle avoidance under such challenging conditions. Additionally, to account for saturation constraints or potential faults in the controller inputs commonly encountered in engineering applications, it implements an asymmetric auxiliary control system. Furthermore, the Lyapunov stability theorem is utilized to ensure the stability of both the formation control and obstacle avoidance algorithms for multiple USVs. Finally, the effectiveness of the proposed algorithm is validated through simulations.

Exploring the late maturation of an intrinsic episodic memory network: A resting-state fMRI study

Previous research suggests that episodic memory relies on functional neural networks,which are present even in the absence of an explicit task. The regions that integrate.these networks and the developmental changes in intrinsic functional connectivity.remain elusive. In the present study, we outlined an intrinsic episodic memory network.(iEMN) based on a systematic selection of functional connectivity studies, and.inspected network differences in resting-state fMRI between adolescents (13–17 years.old) and adults (23–27 years old) from the publicly available NKI-Rockland Sample.Through a review of brain regions commonly associated with episodic memory.networks, we identified a potential iEMN composed by 14 bilateral ROIs, distributed.across temporal, frontal and parietal lobes. Within this network, we found an increase.in resting-state connectivity from adolescents to adults between the right temporal pole.and two regions in the right lateral prefrontal cortex. We argue that the coordination of.these brain regions, connecting areas of semantic processing and areas of controlled.retrieval, arises as an important feature towards the full maturation of the episodic.memory system. The findings add to evidence suggesting that adolescence is a key.period in memory development and highlights the role of intrinsic functional.connectivity in such development.

Influential mechanism of water occurrence states of waste-activated sludge: Specifically focusing on the pore characteristics dominated by cation-organic interactions

The solid pore characteristics are commonly considered as the important influential factors on waste-activated sludge (WAS) dewaterability, and should be related to the cohesive force of bio-flocs dominated by cation-organic interactions at solid-water interface. This study aimed to establish an approach for regulating the solid pore structure of WAS by cationic regulation. The influential mechanism of WAS dewaterability was accordingly explored from the perspective of the pore characteristics dominated by cation-organic interactions. Primarily, with the gradient removal or addition of bivalent cations, the varying pore structure of WAS flocs was tracked by in-situ synchrotron X-ray computed microtomography imaging technique (CMT). The three-dimensional visual model was established to quantify the pore structure parameters of WAS flocs. Following the visualization analysis, the artificial intelligence means, the gradient-weighted class activation mapping (Grad CAM) of three-dimensional convolutional neural network (3D-CNN), was applied for the first time to explore the linkages among solid surface properties, solid pore structure, water occurrence states and sludge dewaterability. It was found that the number and volume of isolated pores jointly determined the mobility and the fractions of vicinal water and interstitial water (p-value ≤ 0.02); also, the decreasing polar or acid-based interfacial free energy with the cationic addition was accompanied with the decreasing isolated pore mean-volume (Pearson coefficient=-0.77, p-value < 0.01). These results indicated that the pore structure characteristics determined the water occurrence states, but the solid porosity strongly depended on the interfacial properties. Accordingly, the molecular docking was applied to explore the interfacial reaction mechanism between Ca2+/Mg2+ and solid compositions in terms of complexation sites, molecular dynamics and free energy calculations. As a result, how the cation-organic interactions affected the pore characteristics through solid surface modification could be clarified, which is expected to serve as theoretical foundation for the development of novel sludge conditioning technologies, i.e., more efforts should be devoted to increasing the dense degree of sludge particles through weakening the hydration repulsion of solid surface.

Dynamic Neural Network Predictive Compensation-Based Point-to-Point Iterative Learning Control With Nonuniform Batch Length.

This article discusses the problem of nonuniform running length in incomplete tracking control, which often occurs in industrial processes due to artificial or environmental changes, such as chemical engineering. It affects the design and application of iterative learning control (ILC) that relies on the strictly repetitive property. Therefore, a dynamic neural network (NN) predictive compensation strategy is proposed under the point-to-point ILC framework. To handle the difficulty of establishing an accurate mechanism model for real process control, the data-driven approach is also introduced. First, applying the iterative dynamic linearization (IDL) technique and radial basis function NN (RBFNN) to construct the iterative dynamic predictive data model (IDPDM) relies on input-output (I/O) signal, and the extended variable is defined by a predictive model to compensate for the incomplete operation length. Then, a learning algorithm based on multiple iteration errors is proposed using an objective function. This learning gain is constantly updated through the NN to adapt to changes in the system. In addition, the composite energy function (CEF) and compression mapping prove that the system is convergent. Finally, two numerical simulation examples are given.

Spatial-temporal pattern of ecosystem services and sustainable development in representative mountainous cities: A case study of Chengdu-Chongqing Urban Agglomeration

The Sustainable Development Goals (SDGs) are essential measure for preserving the balance between human well-being and natural ecosystems. The benefit of preserving ecosystems health play a crucial role in promoting the SDGs by providing stable ecosystem services (ESs). However, the ecological health of mountainous cities is vulnerable, with relative low ecological resilience. To investigate the conflict between ecosystems and sustainable development, this study takes the Chengdu-Chongqing Urban Agglomeration as the study area. The major tasks and results in this study include: (1) using the entropy weighting method and the InVEST model, we combined remote sensing, geographic, and statistical data to quantify three types of SDGs (economic, social, environmental) and four ESs (water yield, soil conservation, habitat quality, carbon storage), and establish a localized sustainable development assessment framework that is applicable to the Chengdu-Chongqing Urban Agglomeration. The results show that from 2014 to 2020, the three types of SDGs exhibited an overall upward trend, with the lowest values occurring in 2016. The gap between different counties has narrowed, but significant regional differences still remain, indicating an unbalanced development status quo. Among the 142 counties, water yield and soil conservation values show a consistent downward trend but occupies significant interannual variations, while habitat quality and carbon storage values increases consistently each year. (2) using Spearman's nonparametric correlation analysis and multiscale geographically weighted regression model to explore the temporal variation and spatial heterogeneity of correlations between county ESs and SDGs. The results showed significant heterogeneity in the spatial trade-offs and synergies between ESs and SDGs, with two pairs of synergies weakening, seven pairs of trade-offs increasing, and the strongest negative correlation between Economic Sustainable Development Goals and habitat quality. (3) we applied the self-organizing mapping neural networks to analyze the spatial clustering characteristics of ESs-SDGs. Based on the spatial clustering effects, we divides the Chengdu-Chongqing Urban Agglomeration into four zones, and different zones have different levels of ESs and SDGs. The targeted strategies should be adopted according to local conditions. This work is of great practical importance in maintaining the stability and sustainable development of the Chengdu-Chongqing Urban Agglomeration ecosystem and provides a scientific reference for the optimal regulation of mountainous cities.

Efficient O-type mapping and routing of large-scale neural networks to torus-based ONoCs

Efficient O-type mapping and routing of large-scale neural networks to torus-based ONoCs

High-performance deep spiking neural networks with 0.3 spikes per neuron

Communication by rare, binary spikes is a key factor for the energy efficiency of biological brains. However, it is harder to train biologically-inspired spiking neural networks than artificial neural networks. This is puzzling given that theoretical results provide exact mapping algorithms from artificial to spiking neural networks with time-to-first-spike coding. In this paper we analyze in theory and simulation the learning dynamics of time-to-first-spike-networks and identify a specific instance of the vanishing-or-exploding gradient problem. While two choices of spiking neural network mappings solve this problem at initialization, only the one with a constant slope of the neuron membrane potential at threshold guarantees the equivalence of the training trajectory between spiking and artificial neural networks with rectified linear units. For specific image classification architectures comprising feed-forward dense or convolutional layers, we demonstrate that deep spiking neural network models can be effectively trained from scratch on MNIST and Fashion-MNIST datasets, or fine-tuned on large-scale datasets, such as CIFAR10, CIFAR100 and PLACES365, to achieve the exact same performance as that of artificial neural networks, surpassing previous spiking neural networks. Our approach accomplishes high-performance classification with less than 0.3 spikes per neuron, lending itself for an energy-efficient implementation. We also show that fine-tuning spiking neural networks with our robust gradient descent algorithm enables their optimization for hardware implementations with low latency and resilience to noise and quantization.

Retraction Note: Simple hemogram to support the decision-making of COVID-19 diagnosis using clusters analysis with self-organizing maps neural network

Retraction Note: Simple hemogram to support the decision-making of COVID-19 diagnosis using clusters analysis with self-organizing maps neural network

Degradation modeling and remaining life prediction for a multi-component system under triple uncertainties

As equipment systems are gradually moving toward large-scale and complex that comprise multi-components, the correlation between components significantly impacts the system’s degradation. Meanwhile, the system degradation process is subject to aleatory uncertainty due to inherent fluctuations, epistemic uncertainty since insufficient information, and measurement uncertainty affected by measurement noise. In this paper, a system-level RUL prediction method based on deep belief network, self-organizing map neural network and uncertain random process is proposed by combining the advantages of deep learning in extracting degraded features with the advantages of statistical-based models in quantifying uncertainty. Firstly, the monitoring variables that characterize the equipment degradation trend are selected based on multi-sensor monitoring information of a single component. Deep belief network is utilized to extract component-level deep hidden degradation features. Subsequently, a system-level health index construction method combining mutual information and self-organizing map neural network is proposed according to correlation analysis between components. Finally, a health index evolution model is constructed to predict system RUL by integrating the advantages of the multi-phase degradation model based on uncertain random processes quantifying multiple uncertainties. The superiority of the proposed method is verified through the C-MAPSS dataset provided by NASA.

An improved self-organizing mapping neural network and its application in fault diagnosis of CNC machine tool servo drive system

An improved self-organizing mapping neural network and its application in fault diagnosis of CNC machine tool servo drive system

CM-Pf deep brain stimulation in polyneuromodulation for epilepsy.

Neuromodulation is a viable option for patients with drug-resistant epilepsies. We reviewed the management of patients with two deep brain neurostimulators. In addition, patients implanted with a device targeting the centromedian-parafascicular (CM-Pf) nuclear complex supplements this report to provide an illustrative case to implantation and programming a patient with three active devices. A narrative review using PubMed and Embase identified patients with drug-resistant epilepsy implanted with more than one neurostimulator was performed. Combinations of vagus nerve stimulation (VNS), deep brain stimulation (DBS), and responsive neurostimulation (RNS) were identified. We provide a background of a newly reported case of an adult with a triple implant eventually responding to CM-Pf DBS as the third implant following suboptimal benefit from VNS and RNS. In review of the literature, dual-device therapy is increasing in reports of use with combinations of VNS, RNS, and DBS to treat patients with drug-resistant epilepsy. We review dual-device implants with thalamic DBS device combinations, functional neural networks, and programming patients with dual devices. CM-Pf is a new target for DBS and has shown a variable response in focal epilepsy. We report the unique case of 28-year-old male with drug-resistant focal epilepsy who experienced a 75% seizure reduction with CM-Pf DBS as his third device after suboptimal responses to VNS and RNS. After 9 months, he also experienced seizure freedom from recurrent focal to bilateral tonic-clonic seizures. No medical or surgical complications or safety issues were encountered. We demonstrate safety and feasibility in an adult combining active VNS, RNS, and CM-Pf DBS. Patients with dual-device therapy who experience a suboptimal response to initial device use at optimized settings should not be considered a neuromodulation "failure." Strategies to combine devices require a working knowledge of brain networks.

Fixed-Wing UAV Pose Estimation Using a Self-Organizing Map and Deep Learning

In many Unmanned Aerial Vehicle (UAV) operations, accurately estimating the UAV’s position and orientation over time is crucial for controlling its trajectory. This is especially important when considering the landing maneuver, where a ground-based camera system can estimate the UAV’s 3D position and orientation. A Red, Green, and Blue (RGB) ground-based monocular approach can be used for this purpose, allowing for more complex algorithms and higher processing power. The proposed method uses a hybrid Artificial Neural Network (ANN) model, incorporating a Kohonen Neural Network (KNN) or Self-Organizing Map (SOM) to identify feature points representing a cluster obtained from a binary image containing the UAV. A Deep Neural Network (DNN) architecture is then used to estimate the actual UAV pose based on a single frame, including translation and orientation. Utilizing the UAV Computer-Aided Design (CAD) model, the network structure can be easily trained using a synthetic dataset, and then fine-tuning can be done to perform transfer learning to deal with real data. The experimental results demonstrate that the system achieves high accuracy, characterized by low errors in UAV pose estimation. This implementation paves the way for automating operational tasks like autonomous landing, which is especially hazardous and prone to failure.