• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.160-162
• /
• 2006

BP 알고리즘은 지역 최소점이나 고원 문제와 같은 수렴 실패문제와 학습 속도가 느리다고 알려져 있다. 이제까지 알려진 BP 알고리즘의 대체 방법들은 수렴 속도와 인자에 따른 수렴의 안정성에 대한 불균형을 해소하는데 치중했다. 기존의 전통적인 BP 알고리즘에서 발생하는 위와 같은 문제를 해결하기 위하여, 본 논문에서는 적은 용량의 저장 공간만을 요구하며 수렴이 빠르고 상대적으로 안정성이 보장되는 알고리즘을 제안한다. 이 방법은 상위연결(upper connections), 은닉층-출력층(hidden to output), 하위연결(lower connections), 입력층-은닉층(input to hidden)에 대해 개별적으로 훈련을 시키는 분리 학습방법을 적용한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.4
• /
• pp.691-698
• /
• 2003

We extend the sue of the method of least square to develop a recursive algorithm for the design of adaptive transversal filters such that, given the least-square estimate of this vector of the filter at iteration n-l, we may compute the updated estimate of this vector at iteration n upon the arrival of new data. We begin the development of the RLS algorithm by reviewing some basic relations that pertain to the method of least squares. Then, by exploiting a relation in matrix algebra known as the matrix inversion lemma, we develop the RLS algorithm. An important feature of the RLS algorithm is that it utilizes information contained in the input data, extending back to the instant of time when the algorithm is initiated. In this paper, we propose new tap weight updated RLS algorithm in adaptive transversal filter with data-recycling buffer structure. We prove that convergence speed of learning curve of RLS algorithm with data-recycling buffer is faster than it of exiting RLS algorithm to mean square error versus iteration number. Also the resulting rate of convergence is typically an order of magnitude faster than the simple LMS algorithm. We show that the number of desired sample is portion to increase to converge the specified value from the three dimension simulation result of mean square error according to the degree of channel amplitude distortion and data-recycle buffer number. This improvement of convergence character in performance, is achieved at the B times of convergence speed of mean square error increase in data recycle buffer number with new proposed RLS algorithm.

• Proceedings of the KIEE Conference
• /
• 2009.07a
• /
• pp.1098_1099
• /
• 2009

In this paper, we present the estimation of wind turbine power generation using Cascade Architectures of Fuzzy Neural Networks(CAFNN). The proposed model uses the wind speed average, the standard deviation and the past output power as input data. The CAFNN identification process uses a 10-min average wind speed with its standard deviation. The method for rule-based fuzzy modeling uses Gaussian membership function. It has three fuzzy variables with three modifiable parameters. The CAFNN's configuration has three Logic Processors(LP) that are constructed cascade architecture and an effective optimization method uses two-level genetic algorithm. First, The CAFNN is trained with one-day average input variables. Once the CAFNN has been trained, test data are used without any update. The main advantage of using CAFNN is having simple structure of system with many input variables. Therefore, The proposed CAFNN technique is useful to predict the wind turbine(WT) power effectively and hence that information will be helpful to decide the control strategy for the WT system operation and application.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.12
• /
• pp.1878-1888
• /
• 1990

In this paper, the input-output interconnection method of the multi-valued signal processing circuit using perfect Shuffle technique and Kronecker product is discussed. Using this method, the design method of circuit of the multi-valued Reed-Muller expansions(MRME) to be used the multi-valued signal processing on finite field GF(p**m) is presented. The proposed input-output interconnection method is shown that the matrix transform is efficient and that the module structure is easy. The circuit design of MRME on FG(p**m) is realized following as` 1) contructing the baisc gates on GF(3) by CMOS T gate, 2) designing the basic cells to be implemented the transform and inverse transform matrix of MRME using these basic gates, 3) interconnecting these cells by the input-output interconnecting method of the multivalued signal processing circuits. Also, the circuit design of the multi-valued signal processing function on GF(3\ulcorner similar to Winograd algorithm of 3x3 array of DFT (discrete fourier transform) is realized by interconnection of Perfect Shuffle technique and Kronecker product. The presented multi-valued signal processing circuits that are simple and regular for wire routing and posses the properties of concurrency and modularity are suitable for VLSI.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.7
• /
• pp.566-577
• /
• 1991

In this paper, the method of image segmentation based on a pyramid of reduced resolution versions of the input input image is persented. In a pyramid structure, two regions (a given pixel and its mother pixels) are compared by the proposed inverse-mother-son variance ratio (IMSVR) method for the detection of an optinal object pixel and are determined whether they are similar enough to be viewed as one region or disparate to be viewed as ditinct regions By the proposed method, an l`timal object pixel has been setectedat some level, it is necessary to retrieve its boundary precisely. Moving down the pyramid to levels of higher resolution is requires. In this paper, the top-sown pyramid traversing algorithm for an image segmentation using a pyrmid structure is presented. Using the computer simulation, the results by the proposed statistical method and object traversing method are investigated for the binary image and the real image at the results of computer simulation, the proposed method of image segmentation based on a pyramid structure seem to have useful properties and deserve consideration as a possible alternative to existing methods of omage segmentation. The computation for the proposed method is required 0 (log n), for an TEX>$n{\times}n$ input image.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.41 no.11
• /
• pp.87-95
• /
• 2004

This paper proposes a new test data compression/decompression method for SoC(Systems-on-a-Chip). The method is based on analyzing the factors that influence test parameters: compression ratio and hardware overhead. To improve compression ratio, the proposed method is based on Modified Statistical Coding (MSC) and Input Reduction (IR) scheme, as well as a novel mapping and reordering algorithm proposed in a preprocessing step. Unlike previous approaches using the CSR architecture, the proposed method is to compress original test data and decompress the compressed test data without the CSR architecture. Therefore, the proposed method leads to better compression ratio with lower hardware overhead than previous works. An experimental comparison on ISCAS '89 benchmark circuits validates the proposed method.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.5
• /
• pp.283-290
• /
• 2004

In this paper, a low-power structure for 512 tap FIR filter in MPEG audio algorithm is proposed. By using CSD(Canonic Signed Digit) form filter coefficients and maximum sharing of input signal sample, it is shown that the number of adders of proposed structure can be minimized. To minimize the number of adders, the proposed structure utilizes the 4 steps of sharing, i.e., common input sharing, linear phase symmetric filter coefficient sharing, block sharing for common input, and common sub-expression sharing. Through Verilog-HDL coding, it is shown that reduction rates in the implementation area and relative power consumption of the proposed structure are 60.3% and 93.9% respectively, comparison to those of the conventional multiplier structure.

• Journal of Satellite, Information and Communications
• /
• v.9 no.4
• /
• pp.32-37
• /
• 2014

The panorama is a good alternative to overcome narrow FOV under study in robot vision, stereo camera and panorama image registration and modeling. The panorama can materialize view with angles wider than human view and provide realistic space which make feeling of being on the scene based on realism. If we use all correspondence, it is too difficult to find strong features and correspondences and assume accurate homography matrix in geographic changes in images as load of calculation increases. Accordingly, we used SURF algorithm to estimate overlapping areas with high similarity by comparing and analyzing the input images' histograms and to detect features. And we solved the problem of input order so we can make panorama by input images without order.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.12
• /
• pp.12-20
• /
• 1993

In a conventional adaptive echo canceller, an ADF(Adaptive Digital Filter) with TDL(Tapped-Delay Line) structure modelling the echo path uses the LMS(Least Mean Square) algorithm to compute the coefficients, and NET detector using energy comparison method prevents the ADF to update the coefficients during the periods of the NET signal presence. The convergence speed of the LMS algorithm depends on the eigenvalue spread ratio of the reference signal and NET detector using the energy comparison method yields poor detection performance if the magnitude of the NET signal is small. This paper presents a new adaptive echo canceller which uses the pre-whitening filter to improve the convergence speed of the LMS algorithm. The pre-whitening filter is realized by using a low-order lattice predictor. Also, a new NET signal detection algorithm is presented, where the start point of the NET signal is detected by computing the cross-correlation coefficient between the primary input and the ADF output while the end point is detected by using the energy comparison method. The simulation results show that the convergence speed of the proposed adaptive echo canceller is faster than that of the conventional echo canceller and the cross-correlation coefficient yields more accurate detection of the start point of the NET signal.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.16 no.3
• /
• pp.1-8
• /
• 1979

This is Part I of a two-part article on Korean character processing by a computer. In part I, the problems in Korean character processing are identified and the theoretical foundation is laid out as a viable solution to them. The one-and two-dimensional syntactic structures of Korean characters are formally defined by means of BNF and " Patternal structure " respectively. Formal discussion of lexical and syntactic algorithms is given for character conversion. This character conversion algorithm is applicable to both input and output. For device-independence and implementation-independence, the concept of " cardinal symbol set " is introduced. We will present a historical survey of Korean character processing and discussion of implementation problems for the above algorithm In Part II.lgorithm In Part II.