Comparison of approaches based on optimization and algebraic iteration for binary tomography

Binary tomography represents a special category of tomographic problems, in which only two values are possible for the sought image pixels. The binary nature of the problems can potentially lead to a significant reduction in the number of view angles required for a satisfactory reconstruction, thusly enabling many interesting applications. However, the limited view angles result in a severely underdetermined system of equations, which is challenging to solve. Various approaches have been proposed to address such a challenge, and two categories of approaches include those based on optimization and those based on algebraic iteration. However, the relative strengths, limitations, and applicable ranges of these approaches have not been clearly defined in the past. Therefore, it is the main objective of this work to conduct a systematic comparison of approaches from each category. This comparison suggested that the approaches based on algebraic iteration offered both superior reconstruction fidelity and computation efficiency at low (two or three) view angles, and these advantages diminished at high view angles. Meanwhile, this work also investigated the application of regularization techniques, the selection of optimal regularization parameter, and the use of a local search technique for binary problems. We expect the results and conclusions reported in this work to provide valuable guidance for the design and development of algorithms for binary tomography problems.     

基于兴趣点多特征融合的物体识别方法

提出一种基于兴趣点多种特征融合的物体识别方法。利用简化的局部二值模式算子去除Harris冗余角点,提取感兴趣区域的3种特征并加权融合特征,在K最近邻(KNN)方法中引进加权因子计算特征距离函数,得到合适的分类器。实验结果表明,该方法能有效提高物体识别的正确率。     

时变时滞NCS控制器的改进设计

建立时滞差分不等式,得到网络化控制系统(NCS)的指数稳定判据,基于该判据设计NCS状态反馈控制器。该设计以线性矩阵不等式(LMI)形式给出,不包含任何自由矩阵,通过Matlab中的LMI工具箱进行求解。数值仿真结果表明,该设计具有计算量小、求解方便、保守性弱的特点。     

一种二进制编码的量子粒子群优化算法

针对离散空间优化问题,给出二进制编码的量子粒子群优化（ＢＱＰＳＯ）算法的设计思路,重新定义粒子的位置矢量和粒子之间的距离,提出了ＢＱＰＳＯ 算法的进化方程．通过泛函分析的方法分析了ＢＱＰＳＯ 算法的收敛性,得出全局收敛的结论,并通过多个测试函数测试了ＢＱＰＳＯ 算法的性能．求解结果验证了算法的优越性．     

不确定切换奇异时滞系统鲁棒指数容许性分析

讨论一类连续时间不确定切换奇异区间时变时滞系统的鲁棒指数容许性问题. 通过定义衰减率依赖李亚普诺夫函数并利用平均驻留时间法, 给出一个时滞区间依赖充分条件保证标称系统正则、无脉冲且均方指数稳定. 同时该准则也被推广至不确定系统. 本文获得的结论为连续时间切换奇异时滞系统的基本问题提供了一个解, 即识别切换信号使得切换奇异时滞系统正则、无脉冲且均方指数稳定. 数值例子说明本文结果的有效性.     

Exponential stability and exponential stabilization of singularly perturbed stochastic systems with time‐varying delay

In this paper, the problems of exponential stability and exponential stabilization for linear singularly perturbed stochastic systems with time‐varying delay are investigated. First, an appropriate Lyapunov functional is introduced to establish an improved delay‐dependent stability criterion. By applying free‐weighting matrix technique and by equivalently eliminating time‐varying delay through the idea of convex combination, a less conservative sufficient condition for exponential stability in mean square is obtained in terms of ε‐dependent linear matrix inequalities (LMIs). It is shown that if this set of LMIs for ε=0 are feasible then the system is exponentially stable in mean square for sufficiently small ε?0. Furthermore, it is shown that if a certain matrix variable in this set of LMIs is chosen to be a special form and the resulting LMIs are feasible for ε=0, then the system is ε‐uniformly exponentially stable for all sufficiently small ε?0. Based on the stability criteria, an ε‐independent state‐feedback controller that stabilizes the system for sufficiently small ε?0 is derived. Finally, numerical examples are presented, which show our results are effective and useful. Copyright © 2010 John Wiley & Sons, Ltd.     

Robust exponential stability conditions for retarded systems with Lipschitz nonlinear stochastic perturbations

This paper investigates robust mean‐square exponential stability of a class of uncertain stochastic state‐delayed systems with Lipschitz nonlinear stochastic perturbation. Based on Lyapunov–Krasovskii functional (LKF) method and free‐weighting matrix technique, some new delay‐dependent stability conditions are established in terms of linear matrix inequalities (LMIs). In order to reduce the conservatism, (1) the delay is divided into several segments, i.e. the delay decomposition method is applied; (2) cross terms estimation is avoided; (3) some information of the cross terms relationships which has not been involved in Reference (IET Control Theory Appl. 2008; 2(11):966–973) is considered. Moreover, from the mathematical point of view, the results obtained by free‐weighting matrix technique can be equivalently re‐formulated by simpler ones without involving any additional free matrix variables. The effectiveness of the method is demonstrated by numerical examples. Copyright © 2010 John Wiley & Sons, Ltd.     

Improved exponential stability criteria for uncertain neutral system with nonlinear parameter perturbations

This paper investigates the problem of robust exponential stability for neutral systems with time-varying delays and nonlinear perturbations. Based on a novel Lyapunov functional approach and linear matrix inequality technique, a new delay-dependent stability condition is derived. Since the model transformation and bounding techniques for cross terms are avoided, the criteria proposed in this paper are less conservative than some previous approaches by using the free-weighting matrices. One numerical example is presented to illustrate the effectiveness of the proposed results.     

关于计算矩阵指数的新思考

本文结合符号计算推出了一种计算有理矩阵指数的新方法,利用该方法可以计算出有理矩阵指数的一类精确表达式．这种方法已实现了程序化,并且可以推广到一般有理矩阵函数的计算．