Applied Mechanics and Materials Vol. 610

Abstract: To assist man who is blind or visually impaired on the navigation, a vision system depending upon the DM3730 as the core is developed, which can provide accurate information regarding the environment surrounding the patients. As one part of the project, the feature detection module must be suitable for an embedded implementation, hardware parallelisms and algorithmic modifications need. The Harris detection is chosen as the main algorithm of this module. The experimental results show that the precision and speed of feature extraction are both better than of the PC platform.

Abstract: Under the precondition of relatively adequate historical sample data of obtainable software cost, the thesis makes comprehensive analysis of the advantages and disadvantages of complementary neural network and vector machines, and attempts to study the software cost combined estimation based on RBF neural network and RVM and to build combined estimation model, then applies the entropy evaluation method to identify the weight coefficient of this combined estimation model, and finally it adopts the data from COCOMO database to verify this combined estimation as well as the rationality and scientificity of this model.

Abstract: Using multi-template processing algorithm, the fingerprint features are accurately collected. Through normalization, make the black and white point contrast of the fingerprint image more obviously, strengthen the ridge line texture. Direction calculating algorithm is based on the grey value of the neighborhood pixels. It can be implemented simply and speedily. Through direction filter, noises can be removed, and the contrast of the fingerprint’s ridge lines and valley lines can be enhanced. After binary converting, all information of the fingerprint is stored with 0 and 1. The effect of thinning is to make the fingerprint image more distinct to extract the fingerprint feature point easily. These steps had been implemented on Altera DE2 board with HDL codes. The experimental results indicate that the multi-template algorithm of fingerprint image processing is correct and practicable.

Abstract: A UCA-Root-MUSIC algorithm for direction-of-arrival (DOA) estimation is proposed in this paper which is based on UCA-RB-MUSIC [1]. The method utilizes not only a unitary transformation matrix different from UCA-RB-MUSIC but also the multi-stage Wiener filter (MSWF) to estimate the signal subspace and the number of sources, so that the new method has lower computational complexity and is more conducive to the real-time implementation. The computer simulation results demonstrate the improvement with the proposed method.

Abstract: The stochastic two-sided assembly line is a special line in which workers perform assembly tasks in both sides of the line and the task time is not deterministic. In the improved teaching-learning-based optimization (ITLBO) algorithm, teacher and learner phases are modified for the discrete problem, a novel merging method is proposed to construct the new evolutionary population and Late Acceptance Hill-Climbing (LAHC) is integrated into the self-learning phase to obtain high quality solutions. The experimental results validate the effectiveness of the proposed algorithm.

Abstract: As increasing in the number of elements and the complexity of their interactions, fault prognostics faces real challenge to predict faults in a complex system. System fault regularly results from the interaction of component faults performing as logical and timing relationships. We use Bayesian Net to evaluate these logical relations. And the other section of system faults concerns about time sequence of those component faults. Thence, this Bayesian method is expanded to Time Bayesian Net in order to solve this kind of problem. Component fault prognostics is the basis, running dates derived from sensors are applied to analyze status of components in real time. Then a traditional Bayesian Net is constructed according to the mechanism and logical structure of the system. Followed by, exploiting the conclusion from interaction analysis of components, this net is built as a Time Bayesian Net. Afterward, Timed Bayesian Net receives inputs from the outcomes of component fault prognostics, and predicts the type of fault and its time of occurrence through Bayes’ rules.

Abstract: To solve the problem that traditional threshold segmentation model is not very robust in skin segmentation under different skin colors and different illuminations, an improved adaptive skin color model is proposed. This model detects the change rate of the skin color pixels by modifying the certain threshold while fixing others, then selects the optimum threshold adaptively. The experimental results show that this algorithm can effectively distinguish skin color regions and background regions, and has strong robustness on light disturbance.

Abstract: Analysis jujube leaf rust pigment content and spectral reflectance correlation study comparing jujube leaf rust pigment content and differential spectral correlation. Hyperspectral characteristic parameters to achieve the southern jujube jujube leaf rust pigments PC1/PC2 and PC1+PC2 content estimation. Using a combination of linear and polynomial fitting method to construct the canopy hyperspectral disease dates Brix content estimation model and test. The probabilistic neural network PNN and SVM classifier SVC applied to hyperspectral estimation model, comparative analysis of model accuracy. The results of the quantitative estimation of disease hyperspectral information dates pigment content in leaves of jujube growing use of high spectral monitoring and impact assessment of disease have high practical value.

Abstract: UCT (Upper confidential bounds on Trees) has been applied quite well as a selection approach in MCTS(Monte Carlo Tree Search) in imperfect information games like poker. By using risk dominance as complementary part of decision method besides payoff dominance, opponent strategies is better characterized as their risk factors, like bluff actions in Texas Hold’em Poker . In this paper, estimation method about the influence of risk factors on computing game equilibrium is provided. A novel algorithm, UCT-risk is proposed as modification about UCT algorithm basing on risk estimation methods. To verify the performance of new algorithm, Texas Hold’em, a popular test-bed for AI research is chosen as the experiment platform. The Agent adopted UCT-risk algorithm performs as well or better as the best previous approaches in experiments. And also it is applied in a poker agent named HITSZ_CS_13 in the 2013 AAAI Computer Poker Competition, which confirms the effectiveness of the UCT-risk provided in this paper.

Abstract: Nowadays there are two kinds of tools which are widely used in uncertain information processing, namely, the fuzzy sets theory and fuzzy roughness sets theory. Besides, how to construct a fusion method for the two kinds of uncertain information systematically has been a focus in both academic and applied fields. In this paper, we put forward the concept of level effect function, and analyze the characteristics of a kind of level effect function. Furthermore, we present the fuzzy roughness degree measurement model based on level effect (FRD-BLE), by combining with the roughness measurement method of level cut set. All the results indicate that FRD-BLE has good structural characteristics and interpretability both in theoretical analysis and practical applications. It can be easily apply the fuzzy processing consciousness into decision-making system with the aid help of FRD-BLE.