This paper focuses on the problem of robust reliable control for a class of uncertain switched systems with disturbances and also with actuator failures among a prespecified subset of actuator. Both state matrix and output matrix contain time-varying norm-bounded parameter uncertainties. The purpose of this problem is to design robust feedback-reliable controllers and a switching law such that the systems remain globally quadratically stable and performance not only when all actuators are operational, but also when some actuators experience failures. A design approach based on LMI(linear matrix inequality) method and convex combination technique is proposed to solve the problem addressed. Finally, simulation results illustrate the validity of designed controllers and the switching law.
In this paper, a specific online judge for automata is designed. System architecture is developed according to it. Model representations, judging strategies and feasibility of automatic judge are also explored. Different judging methods are introduced corresponding to each model, ensures the accuracy and efficiency. The advantages when applied to education are also discussed at the end of this paper.
This paper proposes an improving method for the pixel-wise based image watermarking by using image averaging technique and prediction of tuned watermarked pixels. To reduce the error in the original pixel prediction process caused by the high different values of neighbor pixels in the watermarked image, the technique of image averaging is applied to host image in the embedding process, and the watermarked image in the extraction process. Furthermore, a tuning technique is implemented to reduce the prediction error in the original pixel prediction process. The experimental results show the improved performance in term NC at equivalent quality of watermarked images, compared to the previous methods. The robustness against various types of attack of the proposed method is also evaluate and compared.
For the purpose of taking good use of the diagnosis obtained from medical experts and improving the accuracy of chest X-ray images retrieval, the lung fields are segmented and interested regions are marked off on the basis of chest X-ray images having been processed previously; the Gray Difference Statistics is used to indicate the texture feature of each region. Using the K-nearest neighbor classifier, the texture features are mapped respectively to the standard image classes pre-described by the experts, thus it realizes the semantic annotation of regions in the whole image. This method can not only narrow the semantic gap between the low-level features and the high-level semantics of images effectively, but also has an active effect on improving the efficiency of medical diagnosis.
When mosaicking two adjacent orthophotos, the transition from one orthophoto to the other is cannot be seen. The requirements of such an optimal seamline include high color similarity and high texture similarity. In this paper, a novel method integrated snakes and bresenham is presented. The calculation of photometric energy in the snake model considers not only the discrete points, but all the points that the snake passes which found by bresenham. This solves the local optimum problem existed in the snake model to some extent. In a hierarchical strategy, a proper seamline is delineated fully automatically. The experiments carried out in this paper show the promising potential of this schema.
Locating edges in images accurately is important to improve the precision of visual measurement. Firstly, from the characteristics of edge gray distribution in actual images, a method is put forward to detect sub-pixel edges, which fits edge gray distribution with hyperbolic tangent function, and gets subpixel edges through least-square method. Then, the accuracy of the algorithm is evaluated, taking measurement error of sizes of high-accuracy checkerboard as criterion. This evaluation combines with camera calibration technique, which considers factors of lens distortion and perspective distortion.
Face detection is a challenging problem in the field of computer vision and a key component of machine intelligence. Viola and Jones put forward a face detection framework in 2001 that can process images extremely rapidly while achieving high detection rates. This paper extends the framework to handle rotated faces through a simple improvement. The method is to rotate the detected sub-windows by 30 degree clockwise and counterclockwise within the image plane. Experimental results show that the proposed algorithm has a good detection rate and reduces the false positives regardless of the rotation of the faces.
Currently the ASIC applications such as multimedia processing require shorter time-to-market and higher performance. Furthermore, with the IC manufacturing technology developing continually, from transistor level to logic gate level, the size of standard cells in digital circuits is increasing correspondingly. New methodology should be proposed to meet the need for shorter time-to-market IC industry, and the requirement of advanced IC manufacturing technology. This paper proposed the concept and principle of operator design methodology, then focused on the entropy coding application by the reconfigurable operator and finally gave the synthesis data. The results show that compared with the hardware design at the same cost of resources, this methodology can obtain hardware of suitable performance with regular structure. What is more, it can improve the design speed efficiently.
This paper describes a control system of sensorless brushless DC(BLDC) motor. Firstly, the back-EMF is adopted to detect rotor position. Secondly, the back-EMF is very small in the motor starting process, and it is difficult to obtain rotor position efficiently. A re-setting method of the rotor is proposed in the paper, and current closed loop is used for high-speed and safety in the motor starting process. Finally, a good speed and current double closed loops system is designed. The current commutation and motor speed withdraw problem without the position sensor is discussed in details. The experimental results have shown the validity of the sensorless control system and the accuracy of the detective position signal obtained from a prototype BLDC motor of 400 W rating.
In binary brain-computer interfaces (BCI) based on motor imagery, common spatial pattern (CSP) successfully discriminates two-class EEG data. However, low information transfer rate is an intrinsic drawback of binary BCIs that limits their practical applications. It's essential to extend binary CSP algorithm to multi-class paradigms. In this paper, a new approximate joint diagonalization (AJD) method, named unweighted exhaustive diagonalization with Gauss iterations (UEDGI) is proposed for the extension. The UEDGI based multi-class CSP algorithm is applied to five data sets recorded during motor imagery of left hand, right hand, foot or tongue. The performance of the algorithm is accessed by classification accuracy and convergence speed, and compared with other two multi-class CSP algorithms, one versus one (OVO) and one versus the rest (OVR). Experimental results show that the UEDGI based multi-class CSP performs best in both classification rate and running speed.
Text Classification is an important field of research. There are a number of approaches to classify text documents. However, there is an important challenge to improve the computational efficiency and recall. In this paper, we propose a novel framework to segment Chinese words, generate word vectors, train the corpus and make prediction. Based on the text classification technology, we successfully help the Chinese disabled persons to acquire job opportunities efficiently in real word. The results show that using this method to build the classifier yields better results than traditional methods. We also experimentally show that careful selection of a subset of features to represent the documents can improve the performance of the classifiers.
This paper analyzes the regional ionosphere total electron content (TEC) long-term variations obtained via using GPS measurements. For this purpose,the authors used the TEC data from TEC maps (global maps in IONEX format) over 11 years.TEC at five IGS tracking stations of China was interpolated by a suitable method. Results show that the peak of TEC for different stations approximately occurs during the same period from years 1999 to 2009.In detail,there is also a close relationship between the variation of ionospheric TEC and the spatial size of the region.
During the pressure measurement, the measured value may drift with the changing temperature; therefore a special mathematical model should be established to correct the measurement. Several temperature points were selected when calibrating, and the line fitting with least square method was used for test data of pressure at every temperature point. Between the two adjacent temperature points, the piecewise linear interpolation function was used, thus the original curve was replaced with a polyline. A slicing plane fitting was constructed in a three-dimensional space constructed from the values of pressure, temperature and measurement; a three-dimensional graphics of the slicing plane fitting was built by Matlab. Generated from the actual test data and corrected according to the mathematical model, the results coincided with the facts. This model can also be applied to amend pressure in general application situation.
Printed mathematical formulas edited by different soft wares have some obvious differences. To distinguish it before recognition is beneficial to the formula recognition. Based on the statistical analysis to the characteristics of Latex and Word typesetting, an algorithm is designed to identify Latex and Word layout of printed formulas. The experiment indicates that this algorithm can achieve a relative high accuracy to standard printed documents.
We solve a class of convex multi-objective programming problems using aggregate homotopy method. By the means of aggregate function, we transform a multi-objective programming problem with constrains into a uniobjective programming problem with a single constrain. Combined homotopy interior-point method is used to solve the single-object optimization problem; the minimal weak efficient solution of the multi-objective optimization problem has been obtained by path tracking.
Synthetic Aperture Sonar (SAS) performance is limited by motion errors of the towfish. It is necessary to require the towfish moving in a straight line in order to improve the imaging efficiency by Fast Fourier Transformation (FFT). However, the towfish could hardly move along a straight line because of some factors such as ocean wave, ocean current and tide. All kinds of motion errors could incur phase error in echo signal and make the quality of final image decline. The aim of this work is to lighten the influence of the motion errors in final image. Motion errors of towfish are analyzed in detail and a motion compensation (MOCO) scheme based on Inverse Scaled Fourier Transformation (ISCFT) imaging algorithm is proposed.The simulation trials validate the analysis and the scheme.
