A discrete state-space model for linear image processing

The linear time-discrete state-space model is generalized from single-dimensional time to two-dimensional space. The generalization includes extending certain basic known concepts from one to two dimensions. These concepts include the general response formula, state-transition matrix, Cayley-Hamilton theorem, observability, and controllability.     

一种基于参考模型的GPDSP指令流控自动验证方法

随着科学计算和人工智能算法复杂度的增加,作为硬件设计中的控制中心,指令流控部件的设计面临复杂性和精确性急剧提升的挑战.FT-xDSP是国防科技大学自主研发的一款64位GPDSP处理器,其指令流控部件的设计规模和复杂性大幅增加,使得指令流控部件的验证成为一个突出难题.提出一种基于指令重排参考模型的指令流控自动化验证方法:首先,以指令输入输出关系为主要特征建立流控部件的抽象模型,屏蔽了内部复杂逻辑,在保证分析结果准确性的基础上降低了分析复杂度;其次,通过自动生成带约束的随机测试激励,对参考模型和待测设计结果进行自动化比较分析,在验证代价相当的情况下提升了代码覆盖率和功能覆盖率.实验和实际应用结果表明,该方法能针对指令流控验证中的薄弱点进行定向随机验证,大幅度提升了指令流控部件的验证效率和验证完整性.     

一种基于参考模型的GPDSP指令流控自动验证方法

随着科学计算和人工智能算法复杂度的增加,作为硬件设计中的控制中心,指令流控部件的设计面临复杂性和精确性急剧提升的挑战.FT-xDSP是国防科技大学自主研发的一款64位GPDSP处理器,其指令流控部件的设计规模和复杂性大幅增加,使得指令流控部件的验证成为一个突出难题.提出一种基于指令重排参考模型的指令流控自动化验证方法:首先,以指令输入输出关系为主要特征建立流控部件的抽象模型,屏蔽了内部复杂逻辑,在保证分析结果准确性的基础上降低了分析复杂度;其次,通过自动生成带约束的随机测试激励,对参考模型和待测设计结果进行自动化比较分析,在验证代价相当的情况下提升了代码覆盖率和功能覆盖率.实验和实际应用结果表明,该方法能针对指令流控验证中的薄弱点进行定向随机验证,大幅度提升了指令流控部件的验证效率和验证完整性.     

A two-level model for automatic image annotation

Image automatic annotation is a significant and challenging problem in pattern recognition and computer vision. Current image annotation models almost used all the training images to estimate joint generation probabilities between images and keywords, which would inevitably bring a lot of irrelevant images. To solve the above problem, we propose a hierarchical image annotation model which combines advantages of discriminative model and generative model. In first annotation layer, discriminative model is used to assign topic annotations to unlabeled images, and then relevant image set corresponding to each unlabeled image is obtained. In second annotation layer, we propose a keywords-oriented method to establish links between images and keywords, and then our iterative algorithm is used to expand relevant image sets. Candidate labels will be given higher weights by using our method based on visual keywords. Finally, generative model is used to assign detailed annotations to unlabeled images on expanded relevant image sets. Experiments conducted on Corel 5K datasets verify the effectiveness of our hierarchical image annotation model.     

A state-space description for GPC controllers

Generalized predictive control (GPC)-type control algorithms traditionally derived in the polynomial domain are derived in this paper in the state-space domain, but following the polynomial approach due to Clarke et al. (1987). Relations between the polynomial and state-space parameters are presented. Some possible state-space representations which were used earlier in different publications are discussed. The problem of deriving the GPC algorithm in the state-space domain is solved for the unrestricted case as well as for the case of restricted control and output horizons. Some properties of the state estimate for this problem are presented; in particular, two methods of Kalman filtering—optimal and asymptotic—are proposed. The solution is valid for any possible (minimal or non-minimal) state-space representation. Another approach to this problem is by the ‘dynamic programming method’ and solving the Riccati equation (Bitmead et al. 1990). This approach is also presented in this paper but the method differs from this earlier work and does not require extending the state dimension. Ultimately, certain features of the state-space approach are discussed, such as (a) the opportunity for straightforward analysis of the transient states produced by switching on the regulator, by changing the set-point or by changing the regulator parameters; (b) easy extension to the multidimensional case; and (c) the possibility of introducing nonlinearities into the model     

A general algorithm for determining state-space representations

A new and direct procedure is presented for determining state-space representations of given, time-invariant systems whose dynamical behavior is expressed in a more general, differential operator form. The procedure employs some preliminary polynomial matrix operations, if necessary, in order to “reduce” the given system to an equivalent differential operator form which satisfies four specific conditions. An equivalent state-space representation is then determined in a most direct manner; i.e. the algorithm presented requires only a single matrix inversion. An explicit relationship between the partial state and input of the given system and the state of the equivalent state-space system is also obtained.     

The general state-space model for a two-dimensional linear digital system

The general state-space model for a 2-D linear digital system is presented. A new definition of state-transition matrix is given. Based on the definition, it is easy to calculate the state-transition matrix for any linear digital system. The general response formula for a system follows simply from the definition. A new definition of the characteristic function of a system and a theorem parallel to the Cayley-Hamilton theorem are also given. The presented results apply to any linear causal system.     

A state-space formulation for optimal Hankel-norm approximations

The optimal Hankel-norm approximation problem studied in [1] is reformulated in a state-space setting. The crucial extension theorem is reestablished in this framework. The minimal degree optimal approximation is then derived in terms of state-space parameters     

On a generalized algorithm for the transformation of theinput-output model into the state-space model

A generalised algorithm for transformation from an input-output model into a state-space model, which is also suitable for systems with a nondynamic part, is presented. The algorithm is based on a recently developed fast and accurate method which uses a recursive formula. The derivation of the algorithm is presented and a numerical example is also given. A discrete time invariant linear multivariable and completely observable system given in the canonical state-space form with a nondynamic part are considered     

A framework for state-space estimation with uncertain models

Develops a framework for state-space estimation when the parameters of the underlying linear model are subject to uncertainties. Compared with existing robust filters, the proposed filters perform regularization rather than deregularization. It is shown that, under certain stabilizability and detectability conditions, the steady-state filters are stable and that, for quadratically-stable models, the filters guarantee a bounded error variance. Moreover, the resulting filter structures are similar to various (time- and measurement-update, prediction, and information) forms of the Kalman filter, albeit ones that operate on corrected parameters rather than on the given nominal parameters. Simulation results and comparisons with ℋ_∞ guaranteed-cost, and set-valued state estimation filters are provided     

一个基于移动Agent的分布式入侵检测模型

在分析了入侵检测系统的基本状况和移动代理的特点后,指出了目前入侵检测系统存在的不足,提出了基于移动代理的分布式入侵检测模型(MADIDS)。在这个入侵检测系统中,引入了移动代理技术,使入侵检测系统能够跨平台使用;将基于主机和基于网络的检测结合起来,配置多个检测部件代理,各检测部件代理侧重某一方面的工作;并且将所有配置信息存放在数据库中,实现了数据和处理的分离、数据收集、入侵检测和实时响应的分布化。     

Estimating a state-space model from point process observations

A widely used signal processing paradigm is the state-space model. The state-space model is defined by two equations: an observation equation that describes how the hidden state or latent process is observed and a state equation that defines the evolution of the process through time. Inspired by neurophysiology experiments in which neural spiking activity is induced by an implicit (latent) stimulus, we develop an algorithm to estimate a state-space model observed through point process measurements. We represent the latent process modulating the neural spiking activity as a gaussian autoregressive model driven by an external stimulus. Given the latent process, neural spiking activity is characterized as a general point process defined by its conditional intensity function. We develop an approximate expectation-maximization (EM) algorithm to estimate the unobservable state-space process, its parameters, and the parameters of the point process. The EM algorithm combines a point process recursive nonlinear filter algorithm, the fixed interval smoothing algorithm, and the state-space covariance algorithm to compute the complete data log likelihood efficiently. We use a Kolmogorov-Smirnov test based on the time-rescaling theorem to evaluate agreement between the model and point process data. We illustrate the model with two simulated data examples: an ensemble of Poisson neurons driven by a common stimulus and a single neuron whose conditional intensity function is approximated as a local Bernoulli process.     

Low-resolution color-based visual tracking with state-space model identification

A novel tracking method is proposed to resolve the poor performance of color-based tracker in low-resolution vision. The proposed method integrates vector autoregression (VAR) with a conceptual frame of state-space model (SSM) to achieve an appropriate model that clearly describes the relation between high-resolution tracking results (states) and corresponding low-resolution tracking results (observations). Here, the parameters of SSM are calculated by the maximum likelihood (ML) estimator to optimize the SSM and minimize its model error. By using the Kalman filter, known as an effective filter of SSM, to estimate the states of the tracked object from its incomplete observations, it is observed that the estimated states are closer to their actual values than their observations or estimates by other unoptimized SSMs. Therefore, the proposed method can be used to improve low-resolution tracking results. Moreover, it can decrease computational complexity and save on processing time.     

Method for determining a dynamical state-space model for control of thermal MEMS devices

A method is presented for determining lumped dynamical models of thermal microelectromechanical systems (MEMS) devices for purposes of feedback control. As a case study, an electrothermal actuator is used. The physical properties and a set of assumptions are used to determine the basic structure of the dynamical model, which requires the development of the electrical, thermal, and mechanical dynamics. The importance of temperature-dependent parameters is emphasized for dynamical modeling for purposes of feedback control. To confront temperature dependence in a practical yet effective manner, an average temperature is introduced to preserve the energy balance inside the structure. This allows the development of a practical method that combines structure of the model, through the average body temperature, with finite element analysis (FEA) in novel way to perform system identification and identify the unknown parameters. The result is a lumped dynamical model of a MEMS device that can be used for the design of feedback control systems. We compare computer simulated results using the dynamical model with experimental behavior of the actual device to show that our procedure indeed generates an accurate model. This dynamical model is intended for further synthesis of driving signal and control system but also gives a qualitative insight into the relationship between device's geometry and its behavior. The method enables fast development of the model by conducting relatively few static FEA and is verifiable with dynamic experimental results even when temperature measurements are not available.     

A theoretical model for the automatic generation of tag clouds

This paper presents a new approach to information retrieval from non-structured attributes in databases, which involves the processing of text attributes. To make retrieval more effective, frequent text sequences are extracted and mathematically represented as intermediate forms which permit a clearer and more precise definition of operations on texts. These intermediate forms appear to users in the form of tag clouds to facilitate content identification, exploration, and querying. In this sense, tag cloud visualization is a simple, user-friendly visual interface to data. This paper proposes a theoretical model for the representation of frequent text sequences and their operations as well as a general procedure for generating tag clouds from text attributes in databases. The tag clouds thus obtained were compared with conventional tag clouds composed of single terms. Our study showed that automatically generated multi-term tag clouds provide better results than mono-term tag clouds.     

Regularization networks for inverse problems: A state-space approach

Linear inverse problems with discrete data are equivalent to the estimation of the continuous-time input of a linear dynamical system from samples of its output. The solution obtained by means of regularization theory has the structure of a neural network similar to classical RBF networks. However, the basis functions depend in a nontrivial way on the specific linear operator to be inverted and the adopted regularization strategy. By resorting to the Bayesian interpretation of regularization, we show that such networks can be implemented rigorously and efficiently whenever the linear operator admits a state-space representation. An analytic expression is provided for the basis functions as well as for the entries of the matrix of the linear system used to compute the weights. The results are illustrated through a deconvolution problem where the spontaneous secretory rate of luteinizing hormone (LH) of the hypophisis is reconstructed from measurements of plasma LH concentrations.     

Variational Bayesian learning of nonlinear hidden state-space models for model predictive control

This paper studies the identification and model predictive control in nonlinear hidden state-space models. Nonlinearities are modelled with neural networks and system identification is done with variational Bayesian learning. In addition to the robustness of control, the stochastic approach allows for various control schemes, including combinations of direct and indirect controls, as well as using probabilistic inference for control. We study the noise-robustness, speed, and accuracy of three different control schemes as well as the effect of changing horizon lengths and initialisation methods using a simulated cart–pole system. The simulations indicate that the proposed method is able to find a representation of the system state that makes control easier especially under high noise.