We present a type of single-hidden layer feed-forward wavelet neural networks. First, we give a new and quantitative proof
of the fact that a single-hidden layer wavelet neural network with n + 1 hidden neurons can interpolate n + 1 distinct samples with zero error. Then, without training, we constructed a wavelet neural network Xa(x, A), which can approximately interpolate, with arbitrary precision, any set of distinct data in one or several dimensions. The
given wavelet neural network can uniformly approximate any continuous function of one variable. 相似文献
Support vector machines (SVMs) are the effective machine-learning methods based on the structural risk minimization (SRM) principle, which is an approach to minimize the upper bound risk functional related to the generalization performance. The parameter selection is an important factor that impacts the performance of SVMs. Evolution Strategy with Covariance Matrix Adaptation (CMA-ES) is an evolutionary optimization strategy, which is used to optimize the parameters of SVMs in this paper. Compared with the traditional SVMs, the optimal SVMs using CMA-ES have more accuracy in predicting the Lorenz signal. The industry case illustrates that the proposed method is very successfully in forecasting the short-term fault of large machinery. 相似文献
Consensus problem of high-order integral multi-agent systems under switching directed topology is considered in this study. Depending on whether the agent’s full state is available or not, two distributed protocols are proposed to ensure that states of all agents can be convergent to a same stationary value. In the proposed protocols, the gain vector associated with the agent’s (estimated) state and the gain vector associated with the relative (estimated) states between agents are designed in a sophisticated way. By this particular design, the high-order integral multi-agent system can be transformed into a first-order integral multi-agent system. Also, the convergence of the transformed first-order integral agent’s state indicates the convergence of the original high-order integral agent’s state, if and only if all roots of the polynomial, whose coefficients are the entries of the gain vector associated with the relative (estimated) states between agents, are in the open left-half complex plane. Therefore, many analysis techniques in the first-order integral multi-agent system can be directly borrowed to solve the problems in the high-order integral multi-agent system. Due to this property, it is proved that to reach a consensus, the switching directed topology of multi-agent system is only required to be ‘uniformly jointly quasi-strongly connected’, which seems the mildest connectivity condition in the literature. In addition, the consensus problem of discrete-time high-order integral multi-agent systems is studied. The corresponding consensus protocol and performance analysis are presented. Finally, three simulation examples are provided to show the effectiveness of the proposed approach. 相似文献
The present study investigates a large power thermal bubble micro-ejector with induction heating device. The traditional thermal-bubble ejectors adopted resistors as the heating resources, it can only work with lower power and convey liquid with lower flow rate. Induction heating devices are adopted to replace the resistor for heating liquid in this paper. With this heating method, there is no physical contact between the heating core and the external power supply circuit. The liquid in the chamber of micro-ejector is heated by the induction heating device and changes from liquid phase to gas phase, generating vapor bubbles in the micro chamber of the micro ejector. The bubble expands rapidly and ejects droplets through the nozzle. The prototype of the micro-ejector is fabricated and experiments are carried out. Continuous droplets are ejected out from the nozzle as the applied AC current is 0.6–0.65 A with the power frequency of 100 kHz. The total volume of the continuous droplets is ranging from 18.84 to 49.87 nL, and the corresponding flow rate is about 0.52–1.36 μL/min. Furthermore, this new micro-ejector can be adopted in conveying of micro-scale liquid, the injection of trace drugs and the 3D printing. 相似文献
This paper presents a novel three-dimensional fabrication process of micro-scale shell resonator made of single-crystal diamond (SCD) called micro-cupped resonator (MCR). The key feature of the process is the application of laser ablation, which can contribute to the shaping process of SCD with higher accuracy, better structural symmetry, greater speed, and lower breakage rate. UV laser high precision processing machine has been employed for the ablation of SCD in the present investigation. This paper describes the principle of laser ablation of SCD and analyses the effects of beam parameters on ablation quality. Then MCR with better structural symmetry and less structural imperfection has been successfully fabricated based on optimum parameters by laser ablation. The MCR is promising for emerging applications such as micro rate-integrating gyroscope. 相似文献
Large holes are unavoidably generated in depth image based rendering (DIBR) using a single color image and its associated depth map. Such holes are mainly caused by disocclusion, which occurs around the sharp depth discontinuities in the depth map. We propose a divide-and-conquer hole-filling method which refines the background depth pixels around the sharp depth discontinuities to address the disocclusion problem. Firstly, the disocclusion region is detected according to the degree of depth discontinuity, and the target area is marked as a binary mask. Then, the depth pixels located in the target area are modified by a linear interpolation process, whose pixel values decrease from the foreground depth value to the background depth value. Finally, in order to remove the isolated depth pixels, median filtering is adopted to refine the depth map. In these ways, disocclusion regions in the synthesized view are divided into several small holes after DIBR, and are easily filled by image inpainting. Experimental results demonstrate that the proposed method can effectively improve the quality of the synthesized view subjectively and objectively. 相似文献
In this paper, the problem of quantized H∞ control is investigated for a class of 2-D systems described by Roesser model with missing measurements. The measurement missing of system state is described by a sequence of random variables obeying the Bernoulli distribution. Meanwhile, the state measurements are quantized by logarithmic quantizer before being communicated. By introducing a new 2-D Lyapunov-like function, a sufficient condition is derived to guarantee stochastically stable and H∞ performance of the closed-loop 2-D system, where the method of sector-bounded uncertainties is utilized to deal with quantization error. Based on the condition, the quantized H∞ control can be designed by using linear matrix inequality technique. A simulation example is also given to illustrate the proposed method.
With the rapid development of the vehicle industry, crashworthiness has become a crucial aspect in vehicle body design. In fact, crashworthiness is a multivariable optimization design problem for a vehicle body, regardless of structure or material. However, when crashworthiness involves a large number of design variables, including both material and structure variables, it is more difficult to deal with. In this paper, an integrated design technique for materials and structures of vehicle body under crash safety consideration is suggested. First, a finite element model of the vehicle body is established according to relevant vehicle safety standards. Then, the material parameters of the vehicle body are set as analytical factors for factor screening. Next, significant factors are obtained using a three-level saturated design integrated with multi-index comprehensive balance analysis and the MaxUr(3) method, with an improved evaluation method. These screened material parameters along with the corresponding continuous variables of the structure, are considered as the design variables of the integrated design of the vehicle body. Both the weight and the crashworthiness properties are set as the design objectives. Optimal Latin hypercube sampling and radius basis functions are utilized to construct highly accurate surrogate models. Furthermore, the non-dominated sorting genetic algorithm II is implemented to seek the optimal solutions. Finally, two cases considering the roof module and the frontal module of a vehicle body are analyzed to verify the proposed method. 相似文献
Knowledge about the wetland use of migratory bird species during the annual life circle is very interesting to biologists,
as it is critically important in many decision-making processes such as for conservation site construction and avian influenza
control. The raw data of the habitat areas and the migration routes are usually in large scale and with high complexity when
they are determined by high-tech GPS satellite telemetry. In this paper, we convert these biological problems into computational
studies and introduce efficient algorithms for the data analysis. Our key idea is the concept of hierarchical clustering for
migration habitat localizations, and the notion of association rules for the discovery of migration routes from the scattered
location points in the GIS. One of our clustering results is a tree structure, specially called spatial-tree, which is an
illusive map depicting the breeding and wintering home range of bar-headed geese. A related result to this observation is
an association pattern that reveals a high possibility that bar-headed geese’s potential autumn migration routes are likely
between the breeding sites in the Qinghai Lake, China and the wintering sites in Tibet river valley. Given the susceptibility
of geese to spread H5N1, and on the basis of the chronology and the rates of the bar-headed geese migration movements, we
can conjecture that bar-headed geese play an important role in the spread of the H5N1 virus at a regional scale in Qinghai-Tibetan
Plateau. 相似文献