In the choice process of optimal military commu-nication (MC) alternative, evaluation data mainly come from expert judgments, simulation results and test bed data, and they cannot be directly used in evaluation because of differences in form and attribute; and the MC environment changes rapidly as the operation tempo increasing. It is an important effort to judge the effectiveness robustness of MC alternative, since both the evaluation data and the MC envi-ronment are full of uncertainty. A robustness evaluation method based on multiple data sources and Monte Carlo simluation is proposed with respect to the characteristics of them. Mainly include Belief map as data expression form; Regression relational model built with Support Vector Re-gression (SVR) to acquire simulation data’s confidence with test bed data as training example; Extensive Bayesian Algo-rithm (EBA) to fuse data from multiple sources; Beta distri-bution fitting method for each criterion of each alternative by using the fused results; and calculation of the Probability of Best (PoB) of each alternative through Monte Carlo simu-lation. Take MCE evaluation of a Naval Vessels Fleet as an example, the proposed method is compared with some gen-eral methods. The results indicate that the proposed method helps to obtain relatively conservative alternative and is effective in guaranteeing the robustness.

Actuator failures are inevitable in practice especially in complex systems. The unknown failure may cause instability and catastrophic accidents during operation of control systems. A state feedback control scheme is proposed by using backstepping techniques. Compared with exist results, The uncertainties caused by total failure are seen as the bounded term and an estimator is designed to estimate its upper bound. The stability of closed loop system and output tracking performance can be guaranteed by our control law and corresponding update laws of uncertain parameters.

The determination of membership function is fairly critical to fuzzy decision tree induction. Unfortunately, generally used heuristics, such as SLIQ, show the pathological behavior of the attribute tests at split nodes inclining to select a crisp partition. Hence, for induction of binary fuzzy tree, this paper proposes a method depending on the sensitivity degree of attributes to all classes of training examples to determine the transition region of membership function. The method, properly using the pathological characteristic of common heuristics, overcomes drawbacks of G-FDT algorithm proposed by B. Chandra, and it well remedies defects brought on by the pathological behavior. Moreover, the sensitivity degree based algorithm outperforms G-FDT algorithm in respect to classification accuracy.

In this paper, the existence and exponential stability of anti-periodic solutions for shunting inhibitory cellular neural networks (SICNNs) with continuously distributed delays are considered by constructing suitable Lyapunov fuctions and applying some critial analysis techniques. Our results remove restrictive conditions of the global Lipschitz and bounded conditions of activation functions and new sufficient conditions ensuring the exist-ence and exponential stability of anti-periodic solutions for SICNNs are obtained. Moreover, an example is given to illustrate the feasibility of the conditions in our results.

With the rapidly growing use of digital images in medical archival and communication, image compression technology, especially diagnosis lossless compression technology, plays a more and more important role for medical applications. In this thesis, a novel diagnosis loseless compression algorithm is presented for digital mammography. The mammogram is divided into breast region, pectoral muscle and background using the CAD technology. Then mutiple arbitrary shape ROIs coding framework is used to compress the mammogram in which the breast region and pectoral muscle are compressed losslessly and lossily respectively, and the background can be discarded or compressed lossily as user’s will. Experimental results show that the proposed method offer potential advantage in medical applications of digital mammography compression.

Equivalent Material Model is a staple method that can be used to imitate the strata and the ground surface movement which caused by human’s underground activities, such as coal resources extraction. The precision of indoor imitations are mainly decided by ways of models’ deformation observation. This paper proposes a new method of measuring models after analyzing the status of deformation observation of equivalent material models. In this paper, the industrial measuring system is used to measure deformations of the model. The system includes two main parts: photographic surveying of industry instruments and structure lighting scanning devices. The first part is used to get the coordinates of the target points which are set on the model; the second one is used to scan the model surface for catching the cracks on the model surface. Due to the high accuracy of photographic surveying of industry method, it can meet the need of monitoring imperceptible movements of target points; also the structure lighting scanning has a high precision on scanning the model surface, which can get the very thick points cloud of the model surface. The system is originally used for reverse engineering, and it scarcely used for Equivalent material model until our research used, so there is no mature method on data management. This paper researches the special data procession method for the model, and results show that methods in the paper are suit for the industrial measuring system.

SaaS is a new way to deploy software as a hosted service and accessed over the Internet which means the customers don’t need to maintain the software code and data on their own servers. So it’s more important for SaaS systems to take security issues into account. Access control is a security mechanism that enables an authority to access to certain restricted areas and resources according to the permissions assigned to a user. Several access models have been proposed to realize the access control of single instance systems. However, most of the existing models couldn’t address the following SaaS system problems: (1) role name conflicts (2) cross-level management (3) the isomerism of tenants' access control (4) temporal delegation constraints. This paper describes a hierarchical RBAC model called H-RBAC solves all the four problems of SaaS systems mentioned above. This model addresses the SaaS system access control in both system level and tenant level. It combines the advantages of RBDM and ARBAC97 model and introduces temporal constraints to SaaS access control model. In addition, a practical approach to implement the access control module for SaaS systems based on H-RBAC model is also proposed in this paper.

Analysis on the current research status, this article studies on the dynamic subsidence principles of overburden rock strata during coal mining based on similar material simulation test. Close ranged industrial photogram metric system was introduced to collect data. After coordinate transformation, matching and model amendment, dynamic subsidence curves which can be used to analyze the continuity characteristics of overburden subsidence, changes of boundary angle and displacement angle, volume transferring law from rock to surface, etc. were got. The result is useful in further study of the dynamic rule of overburden strata and enriches mining subsidence principles.