Journal of Ambient Intelligence and Humanized Computing最新文献

Crowds can lead up to severe disasterous consequences resulting in fatalities. Videos obtained through public cameras or captured by drones flying overhead can be processed with artificial intelligence-based crowd analysis systems. Being a hot area of research over the past few years, the goal is not only to identify the presence of crowds but also to predict the probability of crowd-formation in order to issue timely warnings and preventive measures. Such systems will significantly reduce the probablity of the potential disasters. Developing effective systems is a challenging task, especially due to factors such as naturally occuring diverse conditions, variations in people or background pixel areas, noise, behaviors of individuals, relative amounts/distributions/directions of crowd movements, and crowd building reasons. This paper proposes an infrared video processing system based on U-Net convolutional neural network for crowd monitoring in infrared video frames to help estimate the people crowd with normal or abnormal trends. The proposed U-Net architecture aims to efficiently extract crowd features, achieve sufficient people marking-up accuracy, competitively with optimal network configurations in terms of the depth and number of filters to consequently minimise the number of coefficients. For further faster processing, hardware resources/implementation area savings, and lower power, the optimized network coefficients measured are represented in Canonic-Signed Digit with minimal number of nonzero (± 1) digits, minimizing the number of underlying shift-add/subtract operations of all multipliers. The achieved significantly reduced computational cost makes the proposed U-Net effectively suitable for resource-constrained and low power applications.

Sign language recognition is an important social issue to be addressed which can benefit the deaf and hard of hearing community by providing easier and faster communication. Some previous studies on sign language recognition have used complex input modalities and feature extraction methods, limiting their practical applicability. This research aims to compare two custom-made convolutional neural network (CNN) models for recognizing American Sign Language (ASL) letters from A to Z, and determine which model performs better. The proposed models utilize a combination of CNN and Softmax activation function, which are powerful and widely used classification methods in the field of computer vision. The purpose of the proposed study is to compare the performance of two specially created CNN models for identifying 26 distinct hand signals that represent the 26 English alphabets. The study found that Model_2 had better overall performance than Model_1, with an accuracy of 98.44% and F1 score 98.41%. However, the performance of each model varied depending on the specific label, suggesting that the choice of model may depend on the specific use case and the labels of interest. This research contributes to the growing field of sign language recognition using deep learning techniques and highlights the importance of designing custom models.

There have been many connections between medical science and artificial intelligence in recent years. Many problems arise with the integrity of communication. Cardiac arrhythmia, carried out using artificial intelligence methods, is one of the most dangerous diseases in the field of prevention. Topics introduced in artificial intelligence are the automatic selection of balancing and classification algorithms. In this study, metrics for machine learning algorithm selection are presented. The first problem is the problem of choosing the best balancing algorithm to balance the data sets, introduced as triangle rate (TR). The second issue to be studied is selecting the best automatic classification algorithm. The third action was to use a scoring algorithm to predict sinus and non-sinus arrhythmias. The heptagonal reinforcement learning (HRL) achieved results competitive with standard algorithms by combining three types of algorithms. The data used in this study was a 12-lead electrocardiogram (ECG) database of arrhythmias. The number of patients examined in this dataset is 10,646. The HRL algorithm has improved the previous algorithms by 5%, achieving 86% cardiac arrhythmia prediction.

Smart agriculture has become one of the most popular technologies for farmers due to its simplicity, ease of deployment, high efficiency, and low overheads. But due to an exponential increase in smart-farming data generation, it is necessary to design secure storage interfaces, that can be scaled for multiple farms. Existing storage models either showcase high security, or high storage efficiency, but a very few models enhance both these parameter sets. Such models are highly complex, and reduce the scalability when applied to large-scale scenarios. To overcome these limitations, this text proposes design of a highly efficient and secure agriculture-record-storage model via reconfigurable blockchains. The proposed model initially uses a multiple crop pattern prediction system via Binary Cascaded Convolutional Neural Network (BC CNN), and deploys a single chained Proof-of-Trust (PoT) based blockchain, that is tuned w.r.t. context of the farms. The prediction is done via weather conditions and soil types. This assists in identification of different crop types, and selection of high trust miner nodes, that can preserve privacy during communication and storage operations. As the blockchain is scaled, a Grey Wolf Optimization (GWO) based model is deployed, which assists in splitting the underlying chain into multiple sidechains. This split is done based on QoS and Security optimizations, which is estimated via temporal miner performance under different farm types. The GWO Model also assists in estimating long-term and high-capacity storage chains, which can be used for archival operations. Due to which, the proposed model is able to improve mining speed by 9.4%, while reducing the energy consumption by 3.5% for different mining operations. The model also defines an indexing strategy for different shards, which assists in increasing data access speed by 12.8% for long-term data sets. Due to these enhancements, the proposed model is capable of deployment for large-scale scenarios.

The aim of this study is to present a novel application of Levenberg–Marquardt backpropagation (LMB) to investigate numerically the solution of functional differential equations (FDE) arising in quantum calculus models (QCMs). The various types of discrete versions of FDM in QCMs are always found to be stiff to solve due to involvement of delay and to overcome the said difficulty, we proposed intelligent computing platform via LMB networks. In order to generate dataset for LMB networks, firstly, the FDEs in QCMs are converted into recurrence relations, then these recurrence systems are solved numerically on a specific input grids in case of both types of FDEs with q-exponential function as well as stable with decreasing behavior characteristics. The training, testing and validation samples based processes are employed to construct LMB networks by exploiting approximation theory on mean square error sense for obtaining the solutions of both types of FDEs. The exhaustive conducted simulation studies for solving FDEs in QCMs via absolute error and mean squared error endorse the accuracy, potential, convergence, stability and worth of proposed technique, which further certified through viable training state parameters, outcomes of error histograms, values of regression/correlation indices.

Abstract

In this work, a wheel-based architecture for 50-gigabit per second next-generation passive optical network stage 2 (50G-NGPON2) represents a promising solution for beyond fifth generation networks. A two-dimensional modified fixed right shifting (2D-MFRS) code is designed and implemented in the proposed architecture to enhance the system capacity and security. The results show that the transmission of 50 Gbps per channel signals over 50–200 km fiber offers high receiver sensitivities of − 17.6 dBm in downlink and − 17.7 dBm in uplink direction with less power penalty of 0.8 dB at the bit error rate of 10^–9. In comparisons with existing optical code division multiple access codes, the proposed architecture using 2D-MFRS code supports upto maximum 260 end subscribers, but also ensures superior performance against the fiber linear and non-linear effects. The simulation results show that the proposed wheel based architecture with 1:128 split ratio drastically improves the fiber reach upto 310 km in uplink and 280 km in downlink direction, compared to other existing passive optical networks (PONs). It is also revealed that the proposed design offers preferable results in terms of high gain and output signal to noise ratio with low noise figure as compared to existing 50 gigabit per second time division multiplexing PON, 50G-NGPON2 and conventional PON. The comparative literature reveals the superiority of proposed design over other existing topologies.

The complex q-rung orthopair fuzzy 2-tuple linguistic set (Cq-ROFTLS), which merges the concepts of complex q-rung orthopair fuzzy sets (Cq-ROFS) and 2-tuple linguistic terms, offers significant advantages in dealing with uncertain and imprecise information during decision-making by effectively representing two-dimensional information within a single set. Notably, the Cq-ROFTLS introduces phase terms that empower experts to express their perspectives flexibly, particularly enhancing its capacity to address periodic elements. To address uncertainty, this approach employs complex values to quantify both membership and non-membership degrees within 2-tuple linguistic environment. Additionally, this research introduces the generalized Maclaurin symmetric mean (MSM) aggregation operator, specifically designed for Cq-ROFTL information. This introduces the Cq-ROFTLMSM and its dual form, the Cq-ROFTL Dual MSM (Cq-ROFTLDMSM), each carrying valuable properties. In cases where the importance of input factors varies, the study proposes the Cq-ROFTL weighted MSM (Cq-ROFTLWMSM) and its dual form, the Cq-ROFTL weighted dual MSM (Cq-ROFTLWDMSM). These operators not only make their debut but also showcase their properties and applications. They flexibly adjust to the significance of inputs, leading to a more refined decision-making process. The methodology extends to address multi-attribute group decision-making (MAGDM) within the Cq-ROFTL framework using the Complex Proportional Assessment (COPRAS) method. The introduction of new aggregation techniques further enhances this approach. A practical illustration involving the selection of the optimal bio-energy production technology (BPT) highlights the real-world effectiveness of the methodology. Through thorough comparisons and a focused exploration of advantages, the study effectively validates the merits of this approach.

In today's highly developed era of information technology, online education is gradually becoming an important teaching mode. Online education provides convenient learning resources and flexible learning methods through online platforms, allowing students to learn according to their own schedule and learning needs. However, compared to traditional education, online education faces some challenges, one of which is how to accurately assess students' learning status. Design an online education student learning status evaluation model based on dual improved neural networks with the aim of improving student learning effectiveness. Using systematic clustering statistical methods to preliminarily analyze the influencing factors of online education students' learning status, and construct an initial evaluation index system; Using the Apriori algorithm to filter the initial indicators, a final online education student learning status evaluation index system is constructed. Using wavelet denoising method to remove noise from evaluation index data, a dual improved radial basis function neural network model is constructed as input. Determine the number of hidden layers in the network using the K-means clustering algorithm, thereby determining the network structure; Based on the optimal network structure, the state transition algorithm is used to adjust the network parameters, and the trained neural network is used for online education student learning state evaluation, outputting the final evaluation result of online education student learning state. The experimental results show that the contribution rate of the model's indicator information reaches 93%, which can accurately evaluate the learning status of online education students based on the optimal model structure and parameters. The above results demonstrate that the constructed model can help teachers and students understand students' learning needs and difficulties in real-time, and provide corresponding teaching support and guidance to promote personalized teaching and improve students' learning experience and outcomes.

An ambient agent may fail to achieve its goal due to the dynamism and nondeterminism of the environment. Based on the concepts of Abilities and Discovery Protocols, we present a context aware failure recovery mechanism for Belief-desire-intention agents. Using a based STRIPS Planner, our approach recovers the failed plan by repairing or replacing faulty actions. Compared to existing works generates the recovery plans dynamically at runtime according to the current context. Indeed, agent abilities are generated and maintained dynamically. To our acknowledge, this is the first time that a dynamic set of abilities has been used in failure recovery.