Advanced Materials Research Vols. 532-533

Abstract: It describes a method to implement frequency offset estimation (FOE) in cell search of the TD-SCDMA system. The received SYNC-DL data is used to calculate frequency offset (FO) before the scrambling code and the Midamble ID is identified. The received Midamble data can be used to calculate FO before Multi-frame synchronization. The FOE with two stages can be used in the same 64chip calculation structure and the hardware resources can be conserved. The simulation results show that the method can control the residual FO within 0.05PPM by smoothing of sub-frame or IIR filtering, which meets the TD-SCDMA system requirements for residual frequency, even though a certain sample bias exist.

Abstract: We propose the concept of algebraic neural network and introduce the algebraic algorithm in the network-training phase, which transforms the complex nonlinear optimization problem to a set of simple algebraic equations and achieves the best result directly. The experimental results show that rub-impact AE source localization problem is well solved by combining the nonlinear dynamic features and neural network, thus to provide cue to rotor rub-impact fault diagnosis and the application prospects and further research values are promising.

Abstract: The automated detection of seizures in EEG is significant for epilepsy monitoring, diagnosis and rehabilitation. In this work, we evaluated the differences between epileptic EEG, interictal EEG and normal EEG by computing their Higuchi Fractal Dimension (HFD) and Approximate Entropy (ApEn). The calculated results show that there are significant differences between epileptic EEG and normal EEG in the variations of HFD and ApEn. HFD and ApEn have been shown to be useful to characterize normal and epileptic brain electrical activities, and the degree of complexity of epileptic EEG is lower than that of normal EEG even during interictal time. Our results could be helpful for interpreting the epileptic brain electrical activity and the normal brain electrical activity, and their neurodynamics.

Abstract: Asynchronous JavaScript and XML (AJAX) has a big resistance to the communication between JavaScript at client side and the service module. AJAX service component is difficult to be integrated with current MVC framework non-invasively. Data exchange system based on structural XML occupies many system resources and transmission bandwidth, so its structure can’t be matched effectively between client and server system. Aiming at these problems, a serialization or deserialization model called EZAjax based on JSON data exchange which can be seamlessly integrated with current mainstream MVC frameworks and can be interfaced by AJAX remote method is designed. With the Inverse of Control (Ioc) container, the JavaScript dynamic stub generation and the JSON serializing model of service module are proposed to implement transparent remote call between JavaScript and the service module inside the container. Application instance of EZAjax is given.

Abstract: Based on a new smoothing function of the well-known nonsmooth FB (Fischer-Burmeis-ter) function, a smoothing Newton-type method for second-order cone programming problems is presented in this paper. The features of this method are following: firstly, the starting point can be chosen arbitrarily; secondly, at each iteration, only one system of linear equations and one line search are performed; finally, global, strong convergence and Q-quadratic convergent rate are obtained. The numerical results demonstrate the effectiveness of the algorithm.

Abstract: The application of distributed multi-sensor information fusion technology in accurate positioning of Underwater Vehicle was introduced in this paper. According to the system structure of Distributed multi-sensor in an AUV “T1”, this article establishes the Kalman filtering mathematical model, accomplishes the fusion algorithm based on Kalman filtering and a numerical simulation. The experimental result shows that the application of fusion algorithm based on Kalman filtering can avoid the limitations of a single sensor, reduce its uncertainty impact and increase the confidence level of data.

Abstract: A numerical method is proposed for solving a sort of constrained continuous minimax problem, in which both the objective function and the constraint functions are continuously differentiable about superior decision variables and are continuous about lower decision variables .Besides,the constraint functions include only superior or lower decision variables．The problem is transformed into unconstrained differentiable problem with the idea of the discrete maximum entropy function and the continuous maximum entropy function and the penalty function method．The basic algorithm is established．The convergence is proofed．Numerical examples are given and show the efficiency and the reliability of the algorithm.

Abstract: Earthquake prediction has always been an extremely important and difficult research topic. A road map was proposed in this paper to capture useful information for earthquake prediction by exploring the time sequence data of groundwater temperature. Firstly, the triangle extreme points and the trend turning points are employed for the piecewise linear representation of the time series data. Then the segmentation is classified and symbolized by slope, and symbol sequence is simplified further according to the simplification rules. Finally, the earthquake catalogue data and the symbol sequence are jointly preprocessed with a new method to form transaction-like data, which then be treated by association analysis to extract earthquake prediction knowledge. The results of experiment show that this processing flow is an effective way to provide valuable information about earthquake prediction.

Abstract: A new class of variational models based on Besov spaces B_1,1^s (s>0 ) and homogeneous Besov space E=B_∞,∞^-1 for image decomposition is proposed. The proposed models can be regarded as generalizations of Aujol-Chambolle model. The associated minimizers of variational problems can be expressed by applying different shrinkage functions which depend on the wavelet scale to each wavelet coefficient. The wavelet based treatment simplifies computation of this class of variational models. Finally, we present numerical results on denoising of both real and remote sensing images.

Abstract: In order to deal with the lack of treatment with high conflict evidence ability in traditional D-S evidence synthesis rules and some defects of improvement method, an adaptive synthesis rule for conflict evidence is presented. Firstly, based on the balance between conjunction and disjunction, the adaptive synthesis rule for conflict evidence is established; Secondly, the probability distribution of the conflict evidences is gathered by Monte Carlo simulation, and then a reasonable fusion balance function between conjunction and disjunction is established; Thirdly, the conflict coefficient is obtained real-time through conjunction; the self-adaptation of evidence synthesis process is realized simultaneously. Finally, the validity of this method is verified through the comparison with other methods.