• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.10
• /
• pp.1234-1243
• /
• 1988

In this paper, a target tracking algorithm using the centroid of moving edges is presented. It aims to avoid the difficulty of image segmentation in case of extracting the centroid from only one frame. The proposed algorithm can more easily segment the target than the conventional one in images with complex background. Moreover, it can track the target well when the target is occluded by an object. The result of applying it to a real-time target tracker is shown to be comparatively good.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2016.10a
• /
• pp.775-778
• /
• 2016

로봇 주행 기술은 전통적인 로봇요소 기술 외에도 여러 기술로 대상 응용서비스에 따라 IT 기술과 적극적인 융합을 통해 다양한 주행방법과 주행성능이 향상되고 있다. 본 논문에서는 대표적인 실내 모바일 로봇인 로봇 청소기를 대상으로 기존의 방법인 적외선과 카메라 방법이 아닌 보통 가정에도 쉽게 존재하는 AP를 이용해 목표를 설정하여 포섭구조 이론을 기반으로 동적인 환경에서도 충전 스테이션 까지 자율 주행이 가능한 로봇 알고리즘을 설계하였다. 그 결과 동적인 환경을 설정하여 로봇이 AP를 찾아가는 것을 확인하였고 주행 경로와 경과 시간을 표로 도출하여 다른 경우를 예측할 수 있게 하였다. 향후 행동 기반 로봇과 다양한 센서를 이용하여 로봇의 위치와 목표점 사이의 최단거리 경로를 구하여 주행하는 것이 목표이다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.2
• /
• pp.237-244
• /
• 2012

This paper studies the design of a two-DOF optimal controller for the strip gauge-tension of cold tandem mill processes, that uses a reference shaping filter and a disturbance observer. First, a mathematical model of the strip gauge and tension system is constructed using the gauge meter equation and Hooke's law, respectively. Next, a two-DOF controller considering of a feedforward controller and a feedback controller is designed. The former is based on the reference shaping filter and the disturbance observer, and the latter is based on the ILQ optimal control algorithm. Finally, it is shown through a computer simulation that the proposed optimal controller is able to improve the strip gauge accuracy and the tension variation more than the conventional MV-AGC controller.

• Journal of KIISE:Software and Applications
• /
• v.31 no.4
• /
• pp.457-466
• /
• 2004

Recommendation is to offer information which fits user's interests and tastes to provide better services and to reduce information overload. It recently draws attention upon Internet users and information providers. The collaborative filtering is one of the widely used methods for recommendation. It recommends an item to a user based on the reference users' preferences for the target item or the target user's preferences for the reference items. In this paper, we propose a neural network based collaborative filtering method. Our method builds a model by learning correlation between users or items using a multi-layer perceptron. We also investigate integration of diverse information to solve the sparsity problem and selecting the reference users or items based on similarity to improve performance. We finally demonstrate that our method outperforms the existing methods through experiments using the EachMovie data.

• Journal of Korea Multimedia Society
• /
• v.11 no.3
• /
• pp.353-363
• /
• 2008

A multi-agent system has a longer learning period and larger state-spaces than a sin91e agent system. In this paper, we suggest a new method to reduce the learning time of Nash Q-learning in a multi-agent environment. We apply Macro-actions to Nash Q-learning to improve the teaming speed. In the Nash Q-teaming scheme, when agents select actions, rewards are accumulated like Macro-actions. In the experiments, we compare Nash Q-learning using Macro-actions with general Nash Q-learning. First, we observed how many times the agents achieve their goals. The results of this experiment show that agents using Nash Q-learning and 4 Macro-actions have 9.46% better performance than Nash Q-learning using only 4 primitive actions. Second, when agents use Macro-actions, Q-values are accumulated 2.6 times more. Finally, agents using Macro-actions select less actions about 44%. As a result, agents select fewer actions and Macro-actions improve the Q-value's update. It the agents' learning speeds improve.

• Journal of Wetlands Research
• /
• v.18 no.1
• /
• pp.16-23
• /
• 2016

The purpose of this study is to analyze quantitatively the inundation causes by applying the prevention of performance objectives using the urban storm water runoff model XP-SWMM. The model was built by using DTM and storm sewer-network with the storm sewer and geo-data of the study area as input-data to assess the current performance of prevention. An analysis of the causes of the inundation by the frequency and the rainfall-duration. As a result, lack of pipe capacity due to flooding, as well as inundation heavier that the backwater rainfall occurs due to the rise of water level of outside. For solve the inundation damage, It is necessary to improvement pipe of capacity lack and installation of a flood control channel.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.3A
• /
• pp.240-248
• /
• 2006

In this paper, we consider an OFDM receiver algorithm design for IEEE 802.11a/p system, which targeting large coverage area while keeping the transmission format unchanged. Particularly, taking into account the inter-symbol interference(ISI) and inter-carrier interference(ICI) that can be induced with large RMS delay spread, we employ channel shortening time-domain equalizer(TEQ) and evaluate the receiver performance in terms of SINR and packet error rate(PER). The preamble defined in IEEE802.11a/p is used to estimated the initial equalizer tap coefficients. Primary purpose of the paper is to give an answer to the question, though partially, whether or not 16-QAM constellation can be used in none line of sight environment at the boundary of a large coverage area. To this end, we first analyze the required TEQ parameters for the target channel environment and then perform simulation for PER performance evaluation in a generic frequency selective fading channel with exponential power-delay profile.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.43 no.12
• /
• pp.1062-1070
• /
• 2015

This paper presents an application of an optimal smoothing filter for moving target tracking problem based on measured noise source. In order to measure distance and velocity for the moving target, a beamforming method is applied to use the noise source by using microphone array. Also a Kalman filter and an optimal smoothing algorithm are adopted to improve accuracy of trajectory estimation by using a Singer target model. The simulation is conducted with a missile dynamics to verify performance of the optimal smoothing filter, and a model rocket is used for experiment environment to compare the trajectory estimation results between the beamforming, the Kalman filter, and the smoother. The Kalman filter results show better tracking performance than the beamforming technique, and the estimation results of the optimal smoother outperform the Kalman filter in terms of trajectory accuracy in the experiment results.

• Journal of Satellite, Information and Communications
• /
• v.11 no.4
• /
• pp.81-88
• /
• 2016

In this paper, we propose an adaptive multi-target estimation algorithm using the characteristics of signals in the IR-UWB(Impulse-Radio Ultra Wideband) radar system, which is attracting attention because it has good transparency, robustness to the indoor environment, and high precision positioning of tens of centimeters. We proposed an algorithm that estimates multiple peaks with the characteristic that the signal reflected by the target has a peak. To verify the performance of these algorithms, multiple targets were placed in front of the radar and the existing technique and the multi - target estimation algorithm were compared. The location of the targets is estimated in real time with one transmitting antenna and one receiving antenna. The number of estimates can be increased compared with the existing peak signal derivation method, and multiple targets can be derived. The conventional technique estimates only one target, which results in a mean square error of 1 while a multi - target estimation algorithm yields a result of about 0.05. The proposed method is expected to be able to apply multiple targets to the estimation and application in one IR-UWB module environment.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.102-104
• /
• 2017

최근 Convolutional neural network (CNN)을 도입하여, SAR 영상의 목표물 인식 알고리즘이 높은 성능을 보여주었다. SAR 영상은 4 종류의 polarization 정보로 구성되어있다. 기계와 신호처리의 비용으로 인하여 일부 데이터는 적은 수의 polarization 정보를 가지고 있다. 따라서 우리는 SAR 영상 data 를 멀티모달 데이터로 해석하였다. 그리고 우리는 이러한 멀티모달 데이터에 잘 작동할 수 있는 콘볼루션 신경망을 제안하였다. 우리는 데이터가 포함하는 모달의 수에 반비례 하도록 scale factor 구성하고 이를 입력 크기조절에 사용하였다. 입력의 크기를 조절하여, 네트워크는 특징맵의 크기를 모달의 수와 상관없이 일정하게 유지할 수 있었다. 또한 제안하는 입력 크기조절 방법은 네트워크의 dead filter 의 수를 감소 시켰고, 이는 네트워크가 자신의 capacity 를 잘 활용한다는 것을 의미한다. 또 제안된 네트워크는 특징맵을 구성할 때 다양한 모달을 활용하였고, 이는 네트워크가 모달간의 상관관계를 학습했다는 것을 의미한다. 그 결과, 제안된 네트워크의 성능은 입력 크기조절이 없는 일반적인 네트워크보다 높은 성능을 보여주었다. 또한 우리는 전이학습의 개념을 이용하여 네트워크를 모달의 수가 많은 데이터부터 차례대로 학습시켰다. 전이학습을 통하여 네트워크가 학습되었을 때, 제안된 네트워크는 특정 모달의 조합 경우만을 위해 학습된 네트워크보다 높은 성능을 보여준다.