Aiming at the characteristics of abnormally slow and diffuse electroencephalogram (EEG) signals of Alzheimer's disease (AD), based on the latent variable analysis (LVA) theory, this study proposed an EEG diagnosis method based on principal component analysis (PCA) and energy feature recognition of potential signals. Firstly, the original electroencephalogram (EEG) signal was reduced into potential signal by principal component analysis (PCA), and then the potential signal was divided into delta, theta, alpha and beta frequency bands by fast Fourier transform (FFT), and then the energy features of the potential signal were extracted. By comparing the results of AD and control group, the distribution of EEG potential signals in AD group was scattered, and the AD rhythm was slowed down in low dimensional space (manifested as the main activity shifting from high frequency to low frequency). Secondly, support vector machine (SVM) classification analysis was performed combined with the energy characteristics of the potential signals, and it was found that the potential signals could effectively identify AD patients, and the recognition effect was better than that of the original EEG signals. In addition, increasing the number of latent signals can improve the SVM classification performance, for example, in the θ-band, the joint features can achieve 94.12% classification accuracy. It is shown that the proposed method preserves the valid information of the original EEG while eliminating the redundant information between EEG channels, which ultimately improves the classification accuracy.
Vehicular edge computing (VEC) is a new paradigm that combines edge computing with vehicular networks to facilitate the efficient processing of data generated by vehicles. In this approach, vehicles offload their computing tasks to nearby roadside units (RSUs) for faster processing. However, VEC faces several challenges in handling offloaded tasks due to the overloading problem on edge servers. As a result, VEC’s performance can deteriorate significantly. To mitigate these challenges, this paper proposes a server capacity planning-based task computation offloading (SCPTCO) approach for vehicular networks (VNs). Server capacity planning is the process of estimating the required capacity of edge servers to handle the offloaded computing tasks from vehicles. Our proposed SCPTCO approach ensures the efficient utilization of edge servers by estimating the required server capacity. Furthermore, we designed a mobility model that considers vehicle movements at various speeds to simulate real-life scenarios. The results of the simulation demonstrated that our proposal performed better than the baseline approaches in all scenarios. It can significantly reduce both service time and task failure rate by 21.3% and 72.2%, respectively, compared to the local edge offloading with random capacity (LEORC) scheme. Moreover, when compared to a local edge offloading with equal capacity (LEOEC) scheme, the reduced rates are around 11.6% and 55.8%, respectively.
A novel non-coherent acoustic underwater multiple-input multiple-output (MIMO) communication system is proposed and analyzed in this article. The proposed system uses space modulation techniques and relies on the time difference of arrival (TDOA) to determine active transmitters. In particular, at each symbol time, only one transmitter is active and transmits an unmodulated data towards the receiver. Via TDOA, the receiver localizes active transmitter and retrieve corresponding data bits. It is revealed in this study that leveraging TDOA with acoustic signals, the receiver can discern the index of the active transmitter precisely and attain very low error probability independent of any channel state information. As well, modulating the index of transmitters, as in space modulation, attains multiplexing gain and enhances the overall data rate. Analytical derivation of the error performance of the proposed system is presented and substantiated through comparison with Monte Carlo simulation results. Reported results reveal that the proposed system is a low complexity technique that is independent of the channel fading distribution, and can be designed to attain zero error probability with proper system dimensions.
A fifth-order gap-coupled resonator loaded ultra-wide band (UWB) filtenna is proposed in this research article. The band-stop Butterworth filter is designed at the ISM frequency of 5.8 GHz and an ultra-wideband circular-shaped microstrip patch is integrated with the extended end of the filter. The filter is designed using the concept of the gap-coupled microstrip stub sections with a metallic via connected at the end. It yields all the necessary capacitive and inductive effects to show its band-stop property at the ISM band. It produces a sharp narrow--band notch at 5.8 GHz and helps the antenna to utilize all the available frequency bands by eliminating the possibility for the antenna to interfere with the 5.8 GHz ISM band.
The Cooperative Spectrum Sensing (CSS) based Cognitive Radio-Internet of Things (CR-IoT) is an advantageous solution to address the spectrum scarcity problem by allowing the unlicensed users to share the available frequency bands. However, this system is vulnerable to malicious attacks due to degrading the detection performance. These attacks consist of the presence of Malicious Users (MUs), who aim to mislead the Fusion Center (FC) with fake or falsification sensing information. This work proposes multiple features-based Machine Learning (ML) like Support Vector Machine (SVM), Naive Bayes (NB), Logistic Regression (LR) and Decision Tree (DT) algorithms, to identify MUs in Orthogonal Frequency Division Multiplexing (OFDM)-based CR-IoT networks. The performance of the proposed scheme is evaluated through simulations. The numerical results show that the proposed scheme achieves an enhanced security where MUs are correctly identified, a better detection gain and an enhanced system sum rate compared with other schemes.
The fixed-point FFT is used widely, particularly in hardware implementations where floating-point arithmetic is not always feasible. To avoid overflow in such implementations, a popular method is to shift values one bit to the right after every FFT stage. Rounding down for this scaling introduces a bias which degrades the SQNR. To avoid this degradation, sometimes convergent rounding is used. In this paper the computationally cheap rounding down is still used, and we observe that the bias is independent of the input signal, so a method is presented to compute the expected final bias ahead of time. During the FFT computation, one subtraction per value is required instead of the more complex convergent rounding, at the cost of only a small additional memory requirement. Apart from more efficient computation, the SQNR is also improved by 0.3 dB compared to using convergent rounding on all tested inputs.
The aim of this work is to study the effect of wind on the performance of unmanned aerial vehicle (UAV) based free-space optical (FSO) communications under a doubly inverted gamma-gamma (IGGG) turbulence channel. To this end, the irradiance probability density function (PDF) and cumulative distribution function (CDF) under random wind fluctuations are derived. Furthermore, considering intensity modulation/direct detection technique, closed-form analytical expressions for outage probability, average bit error rate, and average capacity are derived. The analytical results are verified through Monte-Carlo simulation results.
With the wide application of machine learning, data privacy has been widely concerned. As a distributed machine learning method, federated learning has the characteristic of protecting user data privacy. However, the data between clients is often non-independent and identically distributed (non-iid), which leads to the efficiency of federated learning is greatly reduced. In this paper, we propose a data argumentation method using generative models to solve the problem of data non-iid problem by synthesizing high-quality data at the client. Specifically, each client pre-trains generative models locally according to local data, and then the client synthesizes the missing data according to the distribution of the local data, that is, performs local data argumentation. We take the original non-iid distribution and the normal data argumentation methods in machine learning as two control groups. Multiple experiments show that the data argumentation method using generative model improves the performance of federated learning to some extent.
The low-frequency electric field generated by various electromagnetic effects in the ocean overwrites the useful signal in amplitude and frequency, as a background distraction field for electric field detection. Effective electric field characteristic analysis and prediction methods are helpful to improve the environmental adaptability of the equipment and the ability of target detection. The deep learning method is used to mine the characteristics of underwater electric field in the marine environment to improve the prediction accuracy. In order to explore the relationship between seawater environmental parameters and electric field characteristics, parameters including measure time, seawater conductivity, water depth and wave height are used as independent variables, and the amplitude of the underwater electric field in the marine environment is used as the dependent variable. The relationship model between independent variables and dependent variables are established by deep neural network training to realize the prediction of underwater electric field characteristics in the marine environment. The results of 103 training samples and 2 test samples show that the method can effectively identify the test samples, and can be used for the mining and prediction of underwater electric field characteristics, as well as the detection and identification of weak electric field signals submerged in the marine environment.
To improve the reliability of micro pipetting in automatic pipetting workstations, this paper analyzes and investigates the problems related to abnormal condition recognition and pipetting volume detection during pipetting using deep learning models. Firstly, the YOLOv3 network model is designed to be lightweight, and the DarkNet-53 network, which is the backbone of the network structure for image feature information extraction, is replaced by the MobileNetv3 network to realize the recognition of abnormal conditions such as empty aspiration, air bubbles and blockage during pipetting; then the U-Net network model is built to integrate the CBAM attention mechanism module and the residual module to realize the segmentation of pipetting areas. Then, we construct a U-Net network model incorporating the CBAM attention mechanism module and residual module to segment the pipetting area. The results showed that the model size of the improved YOLOv3 network was reduced by 139.1 MB and the frame rate per second (FPS) was increased by 15.7 frames/s. The improved U-Net network improved the IOU and F1 values for pipetting area segmentation by 8% and 13% respectively.
Integrated Access and Backhauling (IAB) is a viable approach for meeting the unprecedented need for higher data rates of future generations, acting as a cost-effective alternative to dense fiber-wired links. The design of such networks with constraints usually results in an optimization problem of non-convex and combinatorial nature. Under those situations, it is challenging to obtain an optimal strategy for the joint Subchannel Allocation and Power Allocation (SAPA) problem. In this paper, we develop a multi-agent Deep Reinforcement Learning (DeepRL) based framework for joint optimization of power and subchannel allocation in an IAB network to maximize the downlink data rate. SAPA using DDQN (Double Deep Q-Learning Network) can handle computationally expensive problems with huge action spaces associated with multiple users and nodes. Unlike the conventional methods such as game theory, fractional programming, and convex optimization, which in practice demand more and more accurate network information, the multi-agent DeepRL approach requires less environment network information. Simulation results show the proposed scheme's promising performance when compared with baseline (Deep Q-Learning Network and Random) schemes.
In this paper, we investigate the impact of nonzero boresight pointing errors on the performance of free-space optical (FSO) communication systems over Fisher-Snedecor $\mathcal{F}$ atmospheric turbulence channels. To this end, we derive simple and accurate approximations for the probability density function (PDF) and the cumulative distribution function (CDF). The derived PDF and CDF statistics are used to obtain analytical expressions for the outage probability and average bit error rate under two detection techniques, namely intensity modulation direct detection and heterodyne detection techniques. In addition, approximate expressions for the ergodic channel capacity in the high- and low-SNR regimes are derived. The analysis is numerically evaluated and supported by Monte-Carlo simulation results for different turbulence and pointing error scenarios. The results demonstrate that the derived approximate expressions are fairly accurate and provide a precise evaluation of the underlying analyses.
Information Centric Networking (ICN) has been proposed as an encouraging remedy for solving complications of traditional IP based architecture. In this work, we present the design and implementation of an ICN framework in Contiki NG OS, a prominent operating system (OS) for constrained IoT devices that have limited computing and communication capabilities. Elementary end-to-end communication paradigms of Named Data Networking (NDN), inspired from ICN concepts and proposed as one of Future Internet Architectures (FIA), is integrated to Contiki NG OS developed for low-power and constrained Internet of Things (IoT) devices. Our results demonstrate the expediency of using the NDN concept for constrained IoT devices in Contiki NG OS.
This paper introduces the basic principle of neural network and its data processing flow, focusing on the analysis of the network in library information evaluation and evaluation, forecasting and modeling applications. In order to solve the problem that BP Algorithm is easy to fall into local optimization and low training efficiency, a genetic algorithm with good global optimization ability is introduced in the model of book purchasing of University Library based on network neural algorithm, to improve the threshold and weight of BP Neural Network, in order to find the best combination of kernel function type, penalty factor and kernel parameter, and to improve the prediction effect of the model. And then improve the training efficiency of the Algorithm. By optimizing the method of neural network with genetic algorithm, this paper Constructs a book purchasing recommendation model based on genetic neural network to judge the books with high demand from readers, and then generates the recommended purchasing sequence. In the case of limited input, priority procurement needs higher professional books, so that the cost of input is optimized.
Catalysts are substances that can increase the speed of a chemical reaction and are often used in the chemical industry. Palladium is one of the most widely used metal centers in metal-based catalysts, and a lot of palladium complexes have been extensively used in many reactions, particularly in cross-coupling reactions with a carbon−carbon bond formation. All their possible applications as catalysts, along with their uses in biological assays as anticancer agents, make these family of complexes very interesting and highly studied, allowing the modification of the ligands around the metal and the extreme modulation of their properties. Herein we report the synthesis of several palladium cyclometallated compounds with thiosemicarbazone ligands and bis(diphenylphosphino)methane (dppm). Additionally, we evaluate their catalytic activity in a Suzuki−Miyaura cross-coupling reaction, using 4-bromoacetophenone and phenylboronic acid as reagents and following the reaction with 1H-NMR spectroscopy. A final comparison between the catalytic conversions and the complexes allows us to propose the best structure for a catalytic purpose in these conditions.
The optimal layout of asphalt mixing plant is the key problem of expressway construction. This paper studies the basic method of optimal layout of asphalt concrete mixing plant by establishing an integer programming mathematical model and combining the dynamic programming algorithm. the optimal layout of asphalt mixing stations and the supply relationship of asphalt concrete between each section and the asphalt mixing station are obtained through the establishment of integer programming mathematical model and dynamic programming calculation, and the total transportation time of transport vehicles is 8030 min. Combined with the analysis of the layout and transportation scheme of asphalt concrete mixing plant in 7 sections of Linyi Tengzhou expressway, the conclusion can be used as a reference for the application of similar projects.
With all-round changes brought by the Internet in technology, thinking, concept, and mode of the power grid, the production, work, and operation mode of power grid have also undergone brand-new intelligent ecological changes. Mobile informatization has gradually become the "standard configuration" for power grid informatization construction. Given the characteristics of transportation inspection and future development requirements, it is necessary to deeply analyze and apply the integration of "Internet + grid transportation inspection", comprehensively change the traditional transportation inspection management mode and working mode, and build an intelligent mobile Internet system for transportation inspection. At present, the reliability and effectiveness of IOT power transmission and transformation devices are not optimistic. With a large number of device faults and false alarms, long-term operation reliability assessment of the devices of various manufacturers lack effective means and method, and the device availability rate is generally low [1]. The alarm accuracy of the device is difficult to judge, and there is a large amount of invalid data. The data validity cannot be judged in the master station. Monitoring devices of various manufacturers lack automatic monitoring, which brings great trouble to the application of line monitoring device and monitoring information in line operation and maintenance. The reliability system of monitoring device for power transmission and transformation based on transportation and inspection IOT uses Internet area to form integrated, grid and intelligent comprehensive monitoring, diagnosis and service management of reliability and availability of power transmission and transformation state detection device. In this way, it is possible to quickly and conveniently help operation and management personnel adjust the device operation in time, conduct monitoring, alarm analysis and processing, and preventive device maintenance of the power transmission and transformation equipment state monitoring device anytime, anywhere, which provides strong technical support for increasing device availability and enhancing professional support for inspection and control. By business-driven integrated development, it is possible to achieve the goals of improving efficiency, reducing costs, improving services, and creating value.
This paper briefly analyses the noise sources in the use model of micro-electro-mechanical system gyroscope, and introduces an improved design method of adaptive smoothing filter to adjust the filter coefficients by increasing the filter threshold to achieve the noise reduction processing of MEMS gyroscope data. The experimental results show that the improved adaptive smoothing filtering algorithm can effectively suppress the high frequency noise in the gyroscope output signal, improve the measurement accuracy of the gyroscope, and achieve good filtering effect, which helps the stable flight of the quadcopter.
In today's financial markets, more and more people are keen on using machine learning methods to predict the rise and fall of stocks. Successful prediction of ups and downs often represents a reduction in the risk of an investment. In this paper, the work will use machine learning methods to predict the rise and fall of the DJIA index. And this work will add sentiment features and interest rate features in addition to the normal technical features in the selection of data. By comparing the F1-score and accuracy, the result shows that the Gradient boosting machine performs the best among all 5 models. The accuracy of the Gradient boosting model is 62%, so I think our model can be used as a trading algorithm.
Interventional cardiology is a minimally invasive surgery that focuses on treating cardiovascular disorders with catheters, such as coronary artery disease, strokes, peripheral arterial disease, and aortic disease. Ultrasound imaging, also known as echocardiography, is a common imaging technology used to keep track of catheter punctures. Precisely segmenting a medical device during cardiac interventions can increase the procedure's safety and dependability. Deep learning has been widely applied to medical image segmentation tasks. It needs an abundant and representative training dataset to perform well. However, in many cases, the process of collecting and labelling data is costly and time-consuming. In this article, we propose two data augmentation methods, AutoReplace and AutoMove, to solve the problem of data efficiency. We used these methods to train a U-Net model to segment the tiny ablation catheter tip in transesophageal echocardiography into the standard five-chamber views and compared the results with conventional data augmentation methods. The comparison showed that our proposed methods have a better effect on the segmentation performance in this task. The principles of these two methods are simple, and they can be easily applied to other medical images with similar annotations.
