Metamaterials, as a new type of material for manufacturing electromagnetic and electronic devices, have become the focus of many studies. It can be used in many disciplines, such as chemistry, physics, materials science, optoelectronics, semiconductor science, and equipment manufacturing, and are currently a cutting-edge and strategically significant research topic. Electromagnetic metamaterials are a new type of artificial composite electromagnetic material with controllable dielectric constant and magnetic permeability. This paper establishes a theoretical framework for metamaterials, divides electromagnetic wave media into four quadrants using constitutive parameters, and introduces the calculation method of effective magnetic permeability. This paper is based on a square planar spiral structure. In order to reduce the resonant frequency, a quantitative analysis was conducted on the influence of double-layer mirror structure, patch resistance, and different stacking structures on the resonant frequency. The reference resonant frequency was reduced from 9.6MHz to 7.6MHz, providing a solid foundation for subsequent research.
In this paper, the combination of wavelet neural networks (WNN) and sliding mode control (SMC) is proposed and simulated to solve the problem of trajectory-tracking control of a two-link robot manipulator with periodic interference. The difficulties of designing control algorithm are mainly focused on achieving accurate trajectory tracking and good control performance with the guarantee of stability and robustness under uncertain cyclical interference. In order to deal with these issues, WNN is used to approximate the functions of control object and unknown periodic disturbance. In this three-layer neural networks design, a widely used Mexican hat wavelet as an activation function has been applied for hidden-layer neurons. Combined with the SMC theory, the adaptive learning laws of networks parameters are derived in the sense of Lyapunov stability analysis so that the tracking error and convergence of the weight can be guaranteed in this control system. The better effectiveness of proposed SMC and WNN control algorithm is demonstrated by numerical simulation on a two-link robot manipulator, as comparing with that of Gauss Radial Basis Function (GRBF) neural networks.
Diabetes is a serious chronic metabolic disease. In the recent years, more and more studies focus on the use of the non-invasive methods to achieve the blood glucose estimation. More and more consumer technology enterprises focusing on human health are committed to implementing accurate and non-invasive blood glucose algorithm in their products. The near infrared spectroscopy built in the wearable devices is one of the common approaches to achieve the non-invasive blood glucose estimation. However, due to the interference from the external environment, these wearable non-invasive methods yield the low estimation accuracy. Even if it is not medical equipment, as a consumer product, the detection accuracy will also be an important indicator for consumers. To address this issue, this paper employs different models based on different ranges of the blood glucose values for performing the blood glucose estimation. First the photoplethysmograms (PPGs) are acquired and they are denoised via the bit plane singular spectrum analysis (SSA) method. Second, the features are extracted. For the data in the training set, first the features are averaged across the measurements in the feature domain via the optimization approach. Second, the random forest is employed to sort the importance of each feature. Third, the training set is divided into three subsets according to the reference blood glucose values. Fourth, the feature vectors and the corresponding blood glucose values in the same group are employed to build an individual model. Fifth, for each feature, the average of the feature values for all the measurements in the same subset is computed. For the data in the test set, first, the sum of the weighted distances between the test feature values and the average values obtained in the above is computed for each model. Here, the weights are defined based on the importance sorted by the random forest obtained in the above. The model corresponding to the smallest sum is assigned. Finally, the blood glucose value is estimated based on the corresponding model. Compared to the state of arts methods, our proposed method can effectively improve the estimation accuracy. In particular, the mean absolute relative difference (MARD) and the percentage of the data fall in the zone A of the Clarke error grid yielded by our proposed method reaches 12.19%, and 87.0588%, respectively.
This research examines the bipartite synchronization of dynamical network with switching signed topologies, where the nodes are governed by certain nonlinear dynamics, and an event-triggered control strategy is applied by under periodic sampling communications. It is shown that when the topologies switch within a finite set of signed graph, all nodes are guaranteed to achieve bipartite synchronization if the time average of the algebraic connectivity of the corresponding unsigned graph over certain length of time is large enough, and the main theorem details the impacts of nodal dynamics, network structure on synchronization, and gives a criterion for selecting the involved control parameters. Finally, some numerical simulations are presented to show the validity of our theoretical results and the efficiency of the proposed controller.
In this paper, a new method, combined with stationary wavelet transform and Gaussian Radial Basis Function Neural Networks (GRBFNN), is proposed for solving the problem of human gait modelling. Firstly, the hardware system, consisting with MPU6050 sensor, wireless transform module, micro control unit and computer, is designed for collecting the gait signal. Secondly, stationary wavelet transform is applied for decomposing the gait signal with 5 scales. In order to remove the high frequency noise and baseline drift, the coefficients of high frequency and low frequency are set as zero. Thirdly, after wavelet denoising, setting a large enough space to cover the gait signal and establishing lattice points with equal intervals in this space, we take gait signal as input and use lattice points as mapping center in GRBFNN design. Fourthly, the identification equation of continuous dynamical system is rewritten into discrete one, and GRBFNN is used for modelling the dynamical function of gait signal. In order to ensure the stability of iteration, the chosen of gain parameter is proven by the Z transform. Finally, comparing with wavelet neural networks(WNN), the result of test in practice demonstrates the superiority of the proposed method for solving the problem of human gait modelling.
Dynamic multi-objective problems (DMOPs) have aroused extensive attention in recent years. Prediction-based methods have been proven to be effective. However, most existing methods assume the linear relationships between historical solutions. For real-life systems, ignoring the complex nonlinear relationships between historical environments may result in low prediction accuracy. To solve this problem, the echo state network (ESN) based prediction approach is proposed for DMOPs. First, the reservoir of ESN is used to express the input dynamics of the historical solutions to explore the linear or nonlinear relationships among historical solutions. Then, a fractal interpolation technique (FIT) is introduced to enrich the training data while preserving the original time series features as much as possible. The final experimental results show that the designed algorithm can solve the dynamic multi-objective optimization problems effectively.
This paper focuses on a hinterland empty container transportation system involving a road-rail network. We formulate an integer programming model to characterize the empty container allocation and transshipment problem in the road-rail transportation network. The complexity of the studied problem comes from the diversity of inland container transportation modes and the flexibility of empty container allocation between different supply-demand pairs. An effective and efficient Greedy-SPFA (Shortest Path Faster Algorithm) method is proposed to solve this problem through transforming it into a minimum cost flow problem. Computational experiments show that the applied algorithm outperforms the established formulation. And some management inspirations are proposed through sensitivity analysis.
Many researchers and research organisations recognise the value of engaging with consumers with the consistent message being that people who have a personal experience of cancer bring a unique perspective and a wealth of diversity, knowledge, and capabilities to research and health initiatives. In addition, funding bodies almost always insist on researchers demonstrating this perspective has been sought for applications to be considered favourably. However, some hesitation still exists and connecting effectively is often easier said than done. Can a lay person really understand research concepts and provide meaningful input? This session will provide examples of working partnerships and what it takes for individual researchers, research organisations, and consumers to work collaboratively to bring better health outcomes for people affected by cancer.
Efficient and accurate short-term load forecasting (STLF) is significance in modern electricity markets. However, accurate short-term load forecasting is challenging due to the non-stationary power load patterns. In this work, we propose a short-term load forecasting framework based on maximal information coefficient (MIC), moving average filter (MAF) and sample convolution and interactive learning (SCINet), Firstly, MIC is used for feature selection. Secondly, the filtered input features are decomposed using MAF individually. Finally, the data are used in an advanced SCINet for short-term load forecasting. The performance of the proposed method is evaluated using datasets from two different regions of the US electricity market. In addition, we compare the prediction results with support vector regression machines (SVR), long short-term memory networks (LSTM), temporal convolutional networks (TCN), light gradient boosting machine (LightGBM), artificial neural network (ANN), random forest (RF), and sample convolution and interaction networks (SCINet). The proposed model achieves accurate prediction results among all the machine learning models used in this paper.
The end-to-end neural model provides a more robust solution to generate responses than the traditional pipe-line method in the task-oriented dialogue system. However, it is challenging to incorporate the proper knowledge into the gen-erated response, especially when there are substantially related knowledge tuples. This paper proposes a knowledge filter and an attention memory pointer to improve the task-oriented dia-logue model. Specifically, the model uses the knowledge filter to obtain the knowledge tuples most relevant to the keywords of dialog context and builds the knowledge vector. Besides, the task-oriented dialogue model usually needs to copy objects from the correct knowledge tuples to form the question's an-swer. We define an attention memory pointer to help the model choose the correct knowledge tuples. Finally, we conduct ex-periments on the In-Car Assistant dataset. The experimental results show that our model can generate more accurate re-sponses than baseline models in automatic and human evaluations.
Change point detection, as an important technique in artificial intelligence, aims to identify abrupt changes in complex systems. In this paper, we propose a novel gradient-sampling-based approach for change point detection in piecewise linear model. The convergence to a point satisfying the first-order optimality condition is guaranteed. Through extensive numerical experiments, we compare the proposed algorithm with the well known method of Muggeo's segmentation by dynamic programming. By computing the change points on the dataset concerning the relationship between the residential electricity consumption and temperature in Fujian Province, we demonstrate that the proposed algorithm outperforms Muggeo's method. Moreover, when using the change points for power load forecasting, the change points from the proposed algorithm can significantly improve the predictive performance of the Long Short-Term Memory (LSTM) model.
The monitoring of vehicle flow is critical to enable a variety of intelligent transportation systems (ITSs). Traditional vehicle flow monitoring solutions are mainly based on roadside cameras, which may suffer serious performance deterioration in dark environments. In view of this, this paper proposes a Lidar-based vehicle flow monitoring system, which consists three parts: target detection module, vehicle flow counting module and vehicle counting visualization module. Specifically, the target detection module is built based on self-training data and the YOLOv4 network. Vehicle information is collected and preprocessed to speed up the target detection and enhance the accuracy. The vehicles and their positions are then obtained by performing inference with the trained weights for Lidar-based vehicle detection. On this basis, the vehicle counting module applies a multi-object tracking technique to monitor the vehicles which are nearby the detected one. Additionally, the Hungarian algorithm is used to match the surrounding vehicles. In vehicle counting visualization module, we visualize the system output through OpenCv. Finally, we build the system prototype and evaluate the algorithm performance in realistic environments under different night-time traffic situations. The experimental results demonstrate the practicability and robustness of the proposed solutions.
This paper presents a novel pico-watt dual-output voltage reference for implantable medical devices (IMDs). In IMDs, excellent capabilities to reject the noise from power source and work with low power consumption and small active area are critical for voltage references. In the proposed design, a dual-output reference voltage is generated through two sets of 2-transistor (2-T) structure and a shared 4-bit trimming circuit to reduce the effects of process variations. At a typical corner, the proposed circuit generates two reference voltages V ref1 and V ref2 of about 88mV and 228mV and the voltage difference is 0.756mV and 6.546mV respectively from 0 °C to 120 °C, The noise rejection ratio greater than 35dB achieved in the simulation shows that V ref2 has a strong ability to suppress the noise of V ref1 . Therefore, two noise-isolated reference voltages are generated, providing accurate and interference-free reference voltages for noisy and noiseless functional circuits. Furthermore, the power consumption is only 8.52 pW at room temperature and the active area is only 0.0019 mm 2 .
In wastewater treatment process (WWTP), which has nonlinear and dynamic characteristics, it is difficult to realize the tracking control of dissolved oxygen (DO). To solve this problem, the online double heuristic programming (DHP) controller based on leaky echo state network (LESN) is proposed, the controller is named as DHP-LESN for short. Firstly, three Leaky ESNs are used in DHP to produce the control strategy, the system state and the derivatives of evaluation function, respectively. Then, the online gradient algorithm is used to update the output weights of three LESNs. Finally, the performance of the proposed DHP-LESN controller is tested and evaluated on Benchmark Simulation Model 1 (BSM1). The simulation results show that the proposed DHP-LESN controller can achieve better control nerformance than PID controller.
This paper employs the electroencephalogrms (EEGs) to analyze the effect of eating the chocolate on the happiness of the subjects. In particular, the EEGs are acquired by a single channel head band as well as the questionnaires on the happiness are conducted before and after eating the chocolate. Here, each EEG is acquired for 10 minutes. Then, the EEGs are transmitted to the cloud system via a bluetooth module. In the cloud system, the EEGs are first denoised using an ideal lowpass filtering via the discrete Fourier transform approach. Next, different features are extracted from different brain waves localized in different frequency bands. By performing the classification of the EEGs between before eating the chocolate and after eating the chocolate for all the EEGs in the test set, the classification accuracy is employed as the score of the happiness. It is found that our obtained score of the happiness is very close to the score obtained in the questionnaires. This implies that the chocolate can improve the happiness of the subjects and the happiness of the subjects can be reflected by the EEGs.
Diabetes is a chronic metabolic disease. Due to insufficient insulin secretion to control blood glucose or the inability of the body to effectively use insulin, the blood glucose of patients will be higher than the normal value, resulting in various complications, which will seriously affect the health of patients. Real-time monitoring of blood glucose levels is crucial for early screening of high incidence of diabetes, as well as for diagnosis and treatment of patients with diabetes. Is proposed in this paper in the near-infrared (NIR) the application of noninvasive blood glucose level prediction, analyses the statistical characteristics and the relationship between the filter, proposed the concept of some new characteristics of filter, the filter is applied to the analysis of near infrared non-invasive blood glucose estimates, experimental results show that the new features in the machine learning model can improve the effect of the model.
The birth rate of children declined in the developed countries and the number of pets increased at the same time. The market size of pet product increased significantly. The customer is willing to pay more to buy safe and reliable products to their pets. This paper firstly reviews the reason why the birth rate of children declined in the developed countries. Secondly, the market trend of pet food products is discussed. Thirdly, the available international standards for pet foods are reviewed and discussed. Then the major hazardous substances for dogs and cats are listed with the adverse effects. The available analytical pet foods services is summarized and the limitations are discussed. Finally, a research project is proposed.
The proper design of mobility-aware content caching scheme in vehicular networks is the critical expeditor for an efficient Intelligent Transportation System, which enables diverse applications such as content dissemination and the entertainment for commuting passengers. Due to the dynamics characteristic caused by the mobility of vehicles, it is relatively hard to implement accurate caching prediction and collect useful data samples with the traditional method. Using the recent advances in training deep neural networks, we present a deep reinforcement learning framework, namely RL-ResNet-v1, that learns content chunk allocation and makes online chunk compensation policy from high-dimensional inputs corresponding to the characteristics and requirements of users passing by multiple Road Side Units (RSUs) in a Vehicle-to-Infrastructure scenario. The realized online content caching scheme serves to reduce data redundancy in each RSU with finite-capacity while promoting cache hit ratio that should meet chunk sequentially downloaded requirement. Simulation results show our content caching scheme not only achieves more than 20% improvement of the cache hit ratio, and effective cache ratio compared to baseline schemes, but also adapt to the temporal variation of vehicle speed and network bandwidth.
A fast reconstruction method of temperature field based on principal component analysis (PCA) and convolutional autoencoder is proposed in this paper. The two-dimensional temperature field can be quickly reconstructed by inputting the small amounts of sensor data. Principal component analysis is first used to extract key features from high-dimensional prior dataset, and the extracted results are combined with the sensor measurement points information according to the coefficient optimization method to achieve the approximate reconstruction of the temperature field. Then, the reconstruction results are inputted into the convolutional autoencoder model for iterative learning to further reduce the reconstruction error and achieve accurate reconstruction of the temperature field. The effectiveness proposed method has been verified in the boiler combustion simulation experiment, and the experimental results show that the proposed method can reconstruct the two-dimensional temperature field quickly and accurately, which is of great significance to the research of some combustion systems.
Echo state network (ESN) is a kind of recurrent neural network without involving gradient problem. However, the reservoir of ESN often contains hundreds of neurons, whose corresponding high-dimensional state matrix may result in ill-conditioned solution problem. To solve it, the condition number-based evolving ESN (CNEESN) is proposed, whose sub-reservoir is generated by condition number analysis and differential evolution algorithm (DE). Firstly, the influence of condition number on output weight matrix is analyzed. Secondly, the randomly generated singular values are optimized by condition number and DE based optimize strategy. Finally, simulation result on a benchmark dataset has shown the superiority of the proposed CNEESN.
