This article develops a polytopic linear parameter varying (LPV) model and presents a non-fragile H 2 gain-scheduled control for a flexible air-breathing hypersonic vehicle (FAHV). First, the polytopic LPV model of the FAHV can be obtained by using Jacobian linearization and tensor-product (TP) model transformation approach, simulation verification illustrates that the polytopic LPV model captures the local nonlinearities of the original nonlinear system. Second, based on the developed polytopic LPV model, a non-fragile gainscheduled control method is proposed in order to reduce the fragility encountered in controller implementation, a convex optimisation problem with linear matrix inequalities (LMIs) constraints is formulated for designing a velocity and altitude tracking controller, which guarantees H 2 control performance index. Finally, numerical simulations have demonstrated the effectiveness of the proposed approach.
Users’ behavior analysis has become one of the most important research topics, especially in terms of performance optimization, architecture analysis, and system maintenance, due to the rapid growth of search engine users. By adequately performing analysis on log data, researchers and Internet companies can get guidance to better search engines. In this paper, we perform our analysis based on approximately 750 million entries of search requests obtained from log of a real commercial search engine. Several aspects of users’ behavior are studied, including query length, ratio of query refining, recommendation access, and so on. Different information needs may lead to different behaviors, and we address this discussion in this paper. We firmly believe that these analyses would be helpful with respect of improving both effectiveness and efficiency of search engines.
A wideband rectangular patch antenna resonating at 3.5 GHz and 8 GHz frequencies is developed on a flexible substrate, which can be used for wearable applications. The proposed antenna gives a wide impedance bandwidth of 116%, operating from 2.5 GHz to 9.5 GHz, covering most of the ultra-wideband (UWB) operating frequency range. A two-element multiple-input multiple-output (MIMO) system is developed using the proposed antenna, and the mutual coupling between the two antennas for various separations and frequencies is analyzed by using artificial neural networks (ANNs). The neural structure is trained by using different ANN algorithms and a comparative study is made between them. It is shown that, quasi-Newton (QN) and quasi-Newton multi layer perceptron (QN-MLP) algorithms are better in terms of training, testing errors, and correlation coefficient.
The switched reluctance motor (SRM) is applied in various industrial applications due to its profitable advantages. However, the robustness speed of SRM is one of the major drawbacks, which greatly affects the performance of motor. Thus, the aim of this paper is to control the speed of SRM using H-infinity control strategy. This H-infinity control technique is stronger against robustness. In the proposed speed controller, the rotor position of the SRM is applied to the controller. The speed variation of the rotor is determined from the reference speed and applied to the controller as input. Then, the speed variation and the corresponding sensitivity function are determined. The sensitivity function determination is based on the input weight of the controller. The weight adjustment process is repeated until a stable speed condition is achieved. Then, the output of the proposed control technique is compared with the existing control technique and the robustness is analyzed. Here, the existing control techniques considered are proportionalintegral (PI) controller and fuzzy logic controller (FLC)-based PI gain tuning. The proposed control strategy is simulated in MATLAB working platform and the control performance is analyzed.
For an odd prime p, a new sequence family of period p m − 1, size (M − 1)p mr is proposed using multiplicative and additive characters. The upper bound for the maximum magnitude of nontrivial correlations of the sequence family is derived using well-known character sums. The upper bound is shown to be \((r + 1)\sqrt {p^m } + 3\), which meets the Welch bound asymptotically.
In this paper, a fast-convergence distributed support vector machine (FDSVM) algorithm is proposed, aiming at efficiently solving the problem of distributed SVM training. Rather than exchanging information only among immediate neighbor sites, the proposed FDSVM employs a deterministic gossip protocol-based communication policy to accelerate diffusing information around the network, in which each site communicates with others in a flooding and iterative manner. This communication policy significantly reduces the total number of iterations, thus further speeding up the convergence of the algorithm. In addition, the proposed algorithm is proved to converge to the global optimum in finite steps over an arbitrary strongly connected network (SCN). Experiments on various benchmark data sets show that the proposed FDSVM consistently outperforms the related state-of-the-art approach for most networks, especially in the ring network, in terms of the total training time.
By treating information transmission as tiling over the time-frequency plane, we propose a digital signal transmission scheme employing overcomplete frames as modulation pulses. The new scheme can achieve a signaling rate larger than the Nyquist rate. We first analyze the capacity performance of the frame transmission scheme over additive white Gaussian noise (AWGN) channels. It proves that the proposed scheme can achieve the Shannon capacity asymptotically. Next, we design the Gabor frame system parameters in time-frequency dispersive channels. It is shown that the pulses shape and the time-frequency separation should be matched to the channel dispersion parameters to achieve the minimum energy perturbation. Numerical results are presented to verify the theoretical findings.
Over the past few decades, latent variable model (LVM)-based algorithms have attracted considerable attention for the purpose of data dimensionality reduction, which plays an important role in machine learning, pattern recognition, and computer vision. LVM is an effective tool for modeling density of the observed data. It has been used in dimensionality reduction for dealing with the sparse observed samples. In this paper, two LVM-based dimensionality reduction algorithms are presented firstly, i.e., supervised Gaussian process latent variable model and semi-supervised Gaussian process latent variable model. Then, we propose an LVM-based transfer learning model to cope with the case that samples are not independent identically distributed. In the end of each part, experimental results are given to demonstrate the validity of the proposed dimensionality reduction algorithms.
Based on characteristics of alternating current (AC) critical current of high temperature superconducting (HTS) tapes on the frequency, this paper focuses on AC voltage-current (U-I) behaviors of two kinds of high temperature superconducting tapes, by which BSCCO and YBCO carrying different frequency AC currents are tested in liquid nitrogen temperature of 77 K. It is shown that the AC U-I characteristic curves of different tapes consist of two parts, that is, the resistive part and the hysteresis part. Additionally, the n values of the two parts and the relationship between AC critical current and frequency are obtained through experiments. The experimental results agree with calculated ones well, which is useful for the application of HTS tapes to power technology.
Many researchers have worked on the explanation of AdaBoost’s good experimental results in theory. Some work give an upper bound of generalization error in terms of the margin distribution function, while Breiman gave a sharper generalization error bound based on minimum margin. He also developed the arcgv algorithm to maximize the minimum margin, then made the minimum margin larger than AdaBoost. However, its empirical results are even worse than AdaBoost. Therefore, is the minimum margin bound not practical? This paper gives a new concept called Equilibrium margin (Emargin) and proves a new generalization error bound using Emargin, which is always better than minimum margin bound. In addition, we show Emargin is a good indicator of generalization. Then, we conduct experiments showing that the Emargin of AdaBoost is larger than arc-gv, but the generalization error of Ada-Boost is usually better.
The three-dimensional structure of a biomolecule rather than its one-dimensional sequence determines its biological function. At present, the most accurate structures are derived from experimental data measured mainly by two techniques: X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy. Because neither X-ray crystallography nor NMR spectroscopy could directly measure the positions of atoms in a biomolecule, algorithms must be designed to compute atom coordinates from the data. One salient feature of most NMR structure computation algorithms is their reliance on stochastic search to find the lowest energy conformations that satisfy the experimentally-derived geometric restraints. However, neither the correctness of the stochastic search has been established nor the errors in the output structures could be quantified. Though there exist exact algorithms to compute structures from angular restraints, similar algorithms that use distance restraints remain to be developed. An important application of structures is rational drug design where protein-ligand docking plays a critical role. In fact, various docking programs that place a compound into the binding site of a target protein have been used routinely by medicinal chemists for both lead identification and optimization. Unfortunately, despite ongoing methodological advances and some success stories, the performance of current docking algorithms is still data-dependent. These algorithms formulate the docking problem as a match of two sets of feature points. Both the selection of feature points and the search for the best poses with the minimum scores are accomplished through some stochastic search methods. Both the uncertainty in the scoring function and the limited sampling space attained by the stochastic search contribute to their failures. Recently, we have developed two novel docking algorithms: a data-driven docking algorithm and a general docking algorithm that does not rely on experimental data. Our algorithms search the pose space exhaustively with the pose space itself being limited to a set of hierarchical manifolds that represent, respectively, surfaces, curves and points with unique geometric and energetic properties. These algorithms promise to be especially valuable for the docking of fragments and small compounds as well as for virtual screening.
The existing auditory computational models for evaluating speech intelligibility can only account for energetic masking, and the effect of informational masking is rarely described in these models. This study was aimed to make a computational model considering the mechanism of informational masking. Several psychoacoustic experiments were conducted to test the effect of informational masking on speech intelligibility by manipulating the number of masking talker, speech rate, and the similarity of F0 contour between target and masker. The results showed that the speech reception threshold for the target increased as the F0 contours of the masker became more similar to that of the target, suggesting that the difficulty in segregating the target harmonics from the masker harmonics may underlie the informational masking effect. Based on these studies, a new auditory computational model was made by inducing the auditory function of harmonic extraction to the traditional model of speech intelligibility index (SII), named as harmonic extraction (HF) model. The predictions of the HF model are highly consistent with the experimental results.
White noise deconvolution has a wide range of applications including oil seismic exploration, communication, signal processing, and state estimation. Using the Kalman filtering method, the time-varying optimal distributed fusion white noise deconvolution estimator is presented for the multisensor linear discrete time-varying systems. It is derived from the centralized fusion white noise deconvolution estimator so that it is identical to the centralized fuser, i.e., it has the global optimality. It is superior to the existing distributed fusion white noise estimators in the optimality and the complexity of computation. A Monte Carlo simulation for the Bernoulli-Gaussian input white noise shows the effectiveness of the proposed results.
Competitive learning has attracted a significant amount of attention in the past decades in the field of data clustering. In this paper, we will present two works done by our group which address the nonlinearly separable problem suffered by the classical competitive learning clustering algorithms. They are kernel competitive learning (KCL) and graph-based multi-prototype competitive learning (GMPCL), respectively. In KCL, data points are first mapped from the input data space into a high-dimensional kernel space where the nonlinearly separable pattern becomes linear one. Then the classical competitive learning is performed in this kernel space to generate a cluster structure. To realize on-line learning in the kernel space without knowing the explicit kernel mapping, we propose a prototype descriptor, each row of which represents a prototype by the inner products between the prototype and data points as well as the squared length of the prototype. In GMPCL, a graph-based method is employed to produce an initial, coarse clustering. After that, a multi-prototype competitive learning is introduced to refine the coarse clustering and discover clusters of an arbitrary shape. In the multi-prototype competitive learning, to generate cluster boundaries of arbitrary shapes, each cluster is represented by multiple prototypes, whose subregions of the Voronoi diagram together approximately characterize one cluster of an arbitrary shape. Moreover, we introduce some extensions of these two approaches with experiments demonstrating their effectiveness.
It is important to detect interaction effect of multiple genes during certain biological process. In this paper, we proposed, from systems biology perspective, the concept of co-regulated gene module, which consists of genes that are regulated by the same regulator(s). Given a time series gene expression data, a hidden Markov modelbased Bayesian model was developed to calculate the likelihood of the observed data, assuming the co-regulated gene modules are known. We further developed a Gibbs sampling strategy that is integrated with reversible jump Markov chain Monte Carlo to obtain the posterior probabilities of the co-regulated gene modules. Simulation study validated the proposed method. When compared with two existing methods, the proposed approach significantly outperformed the conventional methods.
An attenuated iterative reliability-based majority-logic (AIML) decoding algorithm for low-density parity-check (LDPC) codes is proposed. This novel algorithm is devised based on the orthogonal check-sums of the one-step majority-logic decoding algorithm in conjunction with certain of reliability measures of the received symbols. The computation of reliability measure of the syndrome sum is refined by introducing an attenuation factor. Simulation results show that in binary input-additive white Gaussian noise channel, the AIML algorithm outperforms other popular iterative reliability-based majority-logic decoding algorithms with a slight increase in computational complexity. With maximum iterations five and fifty, the AIML algorithm can achieve almost identical error performance for LDPC codes. No error floor effect can be observed for AIML algorithm down to the bit error rate (BER) of 10 −8 , while error floor appears for sum product algorithm (SPA) around the BER of 10 −7 even with maximum iteration 100. The inherent feature of parallel decoding for AIML algorithm enforces the decoding speed in contrast to those serial decoding schemes, such as weighted bit-flipping algorithm.
Graph-based methods are one of the widely used unsupervised approaches for keyword extraction. In this approach, words are linked according to their co-occurrences within the document. Afterwards, graph-based ranking algorithms are used to rank words and those with the highest scores are selected as keywords. Although graph-based methods are effective for keyword extraction, they rank words merely based on word graph topology. In fact, we have various prior knowledge to identify how likely the words are keywords. The knowledge of words may be frequency-based, position-based, or semantic-based. In this paper, we propose to incorporate prior knowledge with graph-based methods for keyword extraction and investigate the contributions of the prior knowledge. Experiments reveal that prior knowledge can significantly improve the performance of graph-based keyword extraction. Moreover, by combining prior knowledge with neighborhood knowledge, in experiments we achieve the best results compared to previous graph-based methods.
Echo state network (ESN) proposed by Jaeger in 2001 has remarkable capabilities of approximating dynamics for complex systems, such as Mackey-Glass problem. Compared to that of ESN, the scale-free highly-clustered ESN, i.e., SHESN, which state reservoir has both small-world phenomenon and scale-free feature, exhibits even stronger approximation capabilities of dynamics and better echo state property. In this paper, we extend the state reservoir of SHESN using leaky integrator neurons and inhibitory connections, inspired from the advances in neurophysiology. We apply the extended SHESN, called e-SHESN, to the Mackey-Glass prediction problem. The experimental results show that the e-SHESN considerably outperforms the SHESN in prediction capabilities of the Mackey-Glass chaotic time-series. Meanwhile, the interesting complex network characteristic in the state reservoir, including the small-world property and the scale-free feature, remains unchanged. In addition, we unveil that the original SHESN may be unstable in some cases. However, the proposed e-SHESN model is shown to be able to address the flaw through the enhancement of the network stability. Specifically, by using the ridge regression instead of the linear regression, the stability of e-SHESN could be much more largely improved.
