• 한국조명전기설비학회:학술대회논문집
• /
• 한국조명전기설비학회 2009년도 춘계학술대회 논문집
• /
• pp.420-423
• /
• 2009

The conventional fixed gain PI controller is very sensitive to step change of command speed, parameter variation and load disturbances. The precise speed control of interior permanent magnet synchronous motor(IPMSM) drive becomes a complex issue due to nonlinear coupling among its winding currents and the rotor speed as well as the nonlinear electromagnetic developed torque. Therefore, there exists a need to tune the PI controller parameters on-line to ensure optimum drive performance over a wide range of operating conditions. This paper is proposed hybrid intelligent-PI(HIPI) controller of IPMSM drive using adaptive learning mechanism(ALM) and fuzzy neural network(FNN). The proposed controller is developed to ensure accurate speed control of IPMSM drive under system disturbances and estimation of speed using artificial neural network(ANN) controller. The PI controller parameters are optimized by ALM-FNN at all possible operating condition in a closed loop vector control scheme. The validity of the proposed controller is verified by results at different dynamic operating conditions.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2005년도 공동추계학술대회
• /
• pp.516-525
• /
• 2005

Case-based reasoning (CBR) has been widely used in various areas due to its convenience and strength in complex problem solving. Generally, in order to obtain successful results from CBR, effective retrieval of useful prior cases for the given problem is essential. However, designing a good matching and retrieval mechanism for CBR systems is still a controversial research issue. Most prior studies have tried to optimize the weights of the features or selection process of appropriate instances. But, these approaches have been performed independently until now. Simultaneous optimization of these components may lead to better performance than in naive models. In particular, there have been few attempts to simultaneously optimize the weight of the features and selection of the instances for CBR. Here we suggest a simultaneous optimization model of these components using a genetic algorithm (GA). We apply it to a customer classification model which utilizes demographic characteristics of customers as inputs to predict their buying behavior for a specific product. Experimental results show that simultaneously optimized CBR may improve the classification accuracy and outperform various optimized models of CBR as well as other classification models including logistic regression, multiple discriminant analysis, artificial neural networks and support vector machines.

• 한국산학기술학회논문지
• /
• 제7권6호
• /
• pp.1160-1166
• /
• 2006

혼잡한 현대의 교통 상황에서 교통질서를 유지하기 위해 차량에 대한 정보를 아는 것은 매우 중요한 일이다. 본 논문은 차량의 정보를 아는데 있어서 가장 중요한 차량 번호판을 인식하는 새로운 기법을 소개한다. 제안하는 기법은 물체를 분류하는데 있어서 다른 방법보다 우수하다고 알려진 SVM을 이용한다. 번호판 영역을 찾는데는 이중분류 SVM을 이용하고 번호판 문자 인식에서는 다중 분류 SVM을 이용한다. 여러 단계의 영상처리 과정과 인식 과정을 거쳐서 실시간에 처리할 수 있는 시스템으로 여러 종류의 차량 번호판에 대한 인식도 가능하게 한다. 제안한 기법을 이용한 실제적 환경에서의 영상과 인식에 대한 실험결과를 통하여 성능을 입증하였다.

• Journal of Communications and Networks
• /
• 제4권3호
• /
• pp.230-245
• /
• 2002

Usually the network-throughput maximization problem for constant-bit-rate (CBR) circuit-switched traffic is posed for a fixed offered load profile. Then choices of routes and of admission control policies are sought to achieve maximum throughput (usually under QoS constraints). However, similarly to the notion of channel “capacity,” it is also of interest to determine the “network capacity;” i.e., for a given network we would like to know the maximum throughput it can deliver (again subject to specified QoS constraints) if the appropriate traffic load is supplied. Thus, in addition to determining routes and admission controls, we would like to specify the vector of offered loads between each source/destination pair that “achieves capacity.” Since the combined problem of choosing all three parameters (i.e., offered load, admission control, and routing) is too complex to address, we consider here only the optimal determination of offered load for given routing and admission control policies. We provide an off-line algorithm, which is based on Lagrangian techniques that perform robustly in this rigorously formulated nonlinear optimization problem with nonlinear constraints. We demonstrate that significant improvement is obtained, as compared with simple uniform loading schemes, and that fairness mechanisms can be incorporated with little loss in overall throughput.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2012년도 학술발표회
• /
• pp.35-35
• /
• 2012

Most grid-based distributed hydrologic models are complex in terms of data requirements, parameter estimation and computational demand. To address these issues, a simple grid-based hydrologic model is developed in a geographic information system (GIS) environment using storage-release concept. The model is named GIS Storage Release Model (GIS-StoRM). The storage-release concept uses the travel time within each cell to compute howmuch water is stored or released to the watershed outlet at each time step. The travel time within each cell is computed by combining the kinematic wave equation with Manning's equation. The input to GIS-StoRM includes geospatial datasets such as radar rainfall data (NEXRAD), land use and digital elevation model (DEM). The structural framework for GIS-StoRM is developed by exploiting geographic features in GIS as hydrologic modeling objects, which store and process geospatial and temporal information for hydrologic modeling. Hydrologic modeling objects developed in this study handle time series, raster and vector data within GIS to: (i) exchange input-output between modeling objects, (ii) extract parameters from GIS data; and (iii) simulate hydrologic processes. Conceptual and structural framework of GIS StoRM including its application to Pleasant Creek watershed in Indiana will be presented.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권2호
• /
• pp.330-343
• /
• 2011

In this paper, we propose two schemes to reduce the complexity of fixed-complexity sphere decoder (FSD) algorithm in the ordering and tree-search stages, respectively, while achieving quasi-ML performance. In the ordering stage, we propose a QR-decomposition-based FSD signal ordering based on the zero-forcing criterion (FSD-ZF-SQRD) that requires only a few number of additional complex flops compared to the unsorted QRD. Also, the proposed ordering algorithm is extended using the minimum mean square error (MMSE) criterion to achieve better performance. In the tree-search stage, we introduce a threshold-based complexity reduction approach for the FSD depending on the reliability of the signal with the largest noise amplification. Numerical results show that in $8{\times}8$ MIMO system, the proposed FSD-ZF-SQRD and FSD-MMSE-SQRD only require 19.5% and 26.3% of the computational efforts required by Hassibi’s scheme, respectively. Moreover, a third threshold vector is outlined which can be used for high order modulation schemes. In $4{\times}4$ MIMO system using 16-QAM and 64-QAM, simulation results show that when the proposed threshold-based approach is employed, FSD requires only 62.86% and 53.67% of its full complexity, respectively.

• 대한수학회지
• /
• 제46권5호
• /
• pp.895-905
• /
• 2009

Let $B^{n+1}$ be the unit ball in the complex vector space $\mathbb{C}^{n+1}$ with the standard Hermitian metric. Let ${\Sigma}^n={\partial}B^{n+1}=S^{2n+1}$ be the boundary sphere with the induced CR structure. Let f : ${\Sigma}^n{\hookrightarrow}{\Sigma}^N$ be a local CR immersion. If N < 3n - 1, the asymptotic vectors of the CR second fundamental form of f at each point form a subspace of the CR(horizontal) tangent space of ${\Sigma}^n$ of codimension at most 1. We study the higher order derivatives of this relation, and we show that a linearly full local CR immersion f : ${\Sigma}^n{\hookrightarrow}{\Sigma}^N$, N $\leq$ 3n-2, can only occur when N = n, 2n, or 2n + 1. As a consequence, it gives an extension of the classification of the rational proper holomorphic maps from $B^{n+1}$ to $B^{2n+2}$ by Hamada to the classification of the rational proper holomorphic maps from $B^{n+1}$ to $B^{3n+1}$.

• Structural Engineering and Mechanics
• /
• 제40권2호
• /
• pp.257-277
• /
• 2011

In this paper the geometrically nonlinear continuum plate finite element model, hitherto not reported in the literature, is developed using the total Lagrange formulation. With the layerwise displacement field of Reddy, nonlinear Green-Lagrange small strain large displacements relations (in the von Karman sense) and linear elastic orthotropic material properties for each lamina, the 3D elasticity equations are reduced to 2D problem and the nonlinear equilibrium integral form is obtained. By performing the linearization on nonlinear integral form and then the discretization on linearized integral form, tangent stiffness matrix is obtained with less manipulation and in more consistent form, compared to the one obtained using laminated element approach. Symmetric tangent stiffness matrixes, together with internal force vector are then utilized in Newton Raphson's method for the numerical solution of nonlinear incremental finite element equilibrium equations. Despite of its complex layer dependent numerical nature, the present model has no shear locking problems, compared to ESL (Equivalent Single Layer) models, or aspect ratio problems, as the 3D finite element may have when analyzing thin plate behavior. The originally coded MATLAB computer program for the finite element solution is used to verify the accuracy of the numerical model, by calculating nonlinear response of plates with different mechanical properties, which are isotropic, orthotropic and anisotropic (cross ply and angle ply), different plate thickness, different boundary conditions and different load direction (unloading/loading). The obtained results are compared with available results from the literature and the linear solutions from the author's previous papers.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권8호
• /
• pp.4300-4314
• /
• 2019

With the development of video-related applications, media content has increased dramatically through applications. There is a substantial amount of near-duplicate videos (NDVs) among Internet videos, thus NDVR is important for eliminating near-duplicates from web video searches. This paper proposes a novel NDVR system that supports large-scale retrieval and contributes to the efficient and accurate retrieval performance. For this, we extracted keyframes from each video at regular intervals and then extracted both commonly used features (LBP and HSV) and new image features from each keyframe. A recent study introduced a new image feature that can provide more robust information than existing features even if there are geometric changes to and complex editing of images. We convert a vector set that consists of the extracted features to binary code through a set of hash functions so that the similarity comparison can be more efficient as similar videos are more likely to map into the same buckets. Lastly, we calculate similarity to search for NDVs; we examine the effectiveness of the NDVR system and compare this against previous NDVR systems using the public video collections CC_WEB_VIDEO. The proposed NDVR system's performance is very promising compared to previous NDVR systems.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권2호
• /
• pp.467-479
• /
• 2022

Working memory (WM), which plays a vital role in daily activities, is a memory system that temporarily stores and processes information when people are engaged in complex cognitive activities. The influence of music on WM has been widely studied. In this work, we conducted a series of n-back memory experiments with different task difficulties and multiple trials on 14 subjects under the condition of no music and Alpha wave leading music. The analysis of behavioral data show that the change of music condition has significant effect on the accuracy and time of memory reaction (p<0.01), both of which are improved after the stimulation of Alpha wave music. Behavioral results also suggest that short-term training has no significant impact on working memory. In the further analysis of electrophysiology (EEG) data recorded in the experiment, auto-regressive (AR) model is employed to extract features, after which an average classification accuracy of 82.9% is achieved with support vector machine (SVM) classifier in distinguishing between before and after WM enhancement. The above findings indicate that Alpha wave leading music can improve WM, and the combination of AR model and SVM classifier is effective in detecting the brain activity changes resulting from music stimulation.