• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.1810-1815
• /
• 2003

Sliding mode control with nonlinear interpolation in the boundary layer is proposed. A modified sigmoid function is used for nonlinear interpolation in the boundary layer and its parameter is tuned by a fuzzy logic controller. The fuzzy logic controller that takes the distance between the system state and the sliding surface as its input guides the choice of parameter of the modified sigmoid function and the parameter is on-line tuned. Owing to the decreased thickness, the proposed method has better tracking performance than the conventional linear interpolation method. To demonstrate its performance, the proposed control algorithm is applied to a simple nonlinear system model.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2009년도 제40회 하계학술대회
• /
• pp.1869_1870
• /
• 2009

본 논문에서는 다항식 기반 Type-2 Fuzzy Neural Networks(T2FNN)를 설계하고 이를 패턴분류 문제에 적용하여 그 성능을 분석한다. T2FNN은 Fuzzy C-Means(FCM)을 Type-2 Fuzzy C-Means로 확장시킨 것이라 할 수 있으며, Input layer, Fuzzyification layer, Inference layer, Deffuzification layer의 4층 네트워크로 구성된다. interval Type-1 퍼지 집합인 후반부의 연결가중치는 Gradient Descent Method를 이용하여 학습한다. 제안된 RBF 신경회로망은 모의데이터와 패턴인식 성능 평가에 많이 사용되는 machine learning 데이터에 적용하여 패턴 분류기로서의 성능을 평가받는다.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 1995년도 추계학술대회발표논문집; 서울대학교, 서울; 30 Sep. 1995
• /
• pp.325-333
• /
• 1995

Back-propagation is the most widely used algorithm for supervised learning in multi-layer feed-forward networks. However, back-propagation is very slow in convergence. In this paper, a new weight initialization method, called rough map initialization, in multi-layer perceptrons is proposed. To overcome the long convergence time, possibly due to the random initialization of the weights of the existing multi-layer perceptrons, the rough map initialization method initialize weights by utilizing relationship of input-output features with singular value decomposition technique. The results of this initialization procedure are compared to random initialization procedure in encoder problems and xor problems.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2004년도 춘계학술대회 논문집
• /
• pp.955-960
• /
• 2004

In this study, three dimensional FE analysis of concrete slab track has been performed in order to develop the realistic design of precast concrete slab track. The precast slab track system including the precast concrete slab panel and the grout layer is modeled using the three dimensional solid element with crack softening effect. The input load is computed from the one dimensional beam element model constituting the rail and several discrete springs. To investigate the effect of the longitudinal connection of slab panels, two different systems-continuous and discrete systems - are modeled. The analytical results show that the stresses of both the slab panel and the grout layer are in the range of linear elastic, and, at the interface between two adjacent panels, the primary stresses of the grout layer of the discrete system are higher than those of the continuous system. However, The overall stress levels of the grout layer are very low relative to the strength of th grout.

• E2M - 전기 전자와 첨단 소재
• /
• 제8권3호
• /
• pp.352-356
• /
• 1995

Native oxide layer on MCT (HgCdTe) has been grown uniformly in H$\_$2/O$\_$2/ electrolyte through anodic oxidation method. It has been determined that anodic oxidation of HgCdTe in H$\_$2/O$\_$2/ electrolyte proceeds immediately with the input of constant currents without any induction time required for anodic oxideation in KOH electrolyte. Oxide layer with the resistivity of 2*10$\^$10/.ohm.cm and the refractive index of 2.1 suggested the possibility of well matching combination layer with ZnS for MCT MIS device. XPS results indicated that the major components of oxide layer grown in H202 solution is TeO$\_$2/ with the possibility of small amounts of CdTeO$\_$3/.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 1998년도 Proceedings of International Symposium on Remote Sensing
• /
• pp.201-206
• /
• 1998

In this paper, we have considered the method for clustering land cover types over the East Asia from AVHRR data. The feature vectors such that maximum NDVI, amplitude of NDVI, mean NDVI, and NDVI threshold are extracted from the 10-day composite by maximum value composite(MVC) for reducing the effect of cloud contaninations. To find the land cover clusters given by the feature vectors, we are adapted the self-organizing feature map(SOFM) clustering which is the mapping of an input vector space of n-dimensions into a one - or two-dimensional grid of output layer. The approach is to find first the clusters by the first layer SOFM and then merge several clusters of the first layer to a large cluster by the second layer SOFM. In experiments, we were used the 8-km AVHRR data for two years(1992-1993) over the East Asia.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권11호
• /
• pp.5360-5376
• /
• 2017

In this paper, we study the cross-layer design for the AMC/HARQ-based wireless networks, and propose a new dynamic transmission-mode selection scheme to improve system spectrum efficiency. In the proposed scheme, dynamic thresholds for transmission-mode selection in each packet transmission and retransmission are jointly designed under the constraint of the overall packet error rate. Comparing with the existing schemes, the proposed scheme is inclined to apply higher modulation order at the first several (re)transmissions, which corresponds to higher-rate transmission modes thus higher average system spectrum efficiency. We also extend the cross-layer design to MIMO (Multi-input Multi-output) communication scenarios. Numerical results show that the proposed new dynamic transmission-mode selection scheme generally achieves higher average spectrum efficiency than the conventional and existing cross-layer design.

• 전자공학회논문지B
• /
• 제31B권12호
• /
• pp.116-123
• /
• 1994

A modified Hybrid learning rule(MHLR) is proposed, which is derived from combining the Back Propagation algorithm that is known as an excellent classifier with modified Hebbian by changing the orginal Hebbian which is a good feature extractor. The network architecture of MHLR is multi-layered neural network. The weights of MHLR are calculated from sum of the weight of BP and the weight of modified Hebbian between input layer and higgen layer and from the weight of BP between gidden layer and output layer. To evaluate the performance, BP, MHLR and the proposed Hybrid learning rule (HLR) are simulated by Monte Carlo method. As the result, MHLR is the best in recognition rate and HLR is the second. In learning speed, HLR and MHLR are much the same, while BP is relatively slow.

• 제어로봇시스템학회논문지
• /
• 제6권1호
• /
• pp.8-14
• /
• 2000

In this paper, three adaptive sliding mode control algorithms, which self-tune both the sliding mode gain and the boundary layer thickness, are proposed. The first algorithm uses a gain adaptation rule is combined with the boundary layer thickness adaptatioin rule to satisfy the sliding condition. In the third algorithm, the computation burden of the second algorithm is reduced further, and therefore no extra cost is required for real-time implementation. Due to the mixed sliding mode gain and the boundary layer thickness adaptation scheme, the tracking error and the chattering of the control input can be reduced greatly.

• 전자공학회논문지B
• /
• 제32B권9호
• /
• pp.1137-1153
• /
• 1995

It is not yet possible to solve the optimal number of neurons in hidden layer at neural networks. However, it has been proposed and proved by experiments that there is a limit in increasing the number of neuron in hidden layer, because too much incrememt will cause instability,local minima and large error. This paper proposes a module neural controller with pattern recognition ability to solve the above trade-off problems and to obtain fast learning convergence speed. The proposed neural controller is composed of several module having Multi-layer Perrceptron(MLP). Each module have the less neurons in hidden layer, because it learns only input patterns having a similar learning directions. Experiments with six joint robot manipulator have shown the effectiveness and the feasibility of the proposed the parallel module neural controller with pattern recognition perceptron.