• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.1
• /
• pp.32-39
• /
• 2004

The color-based vision systems have been used to recognize our team robots, the opponent team robots and a ball in the robot soccer system. The color-based vision systems have the difficulty in that they are very sensitive to color variations brought by brightness changes. In this paper, a neural network trained with data obtained from various illumination conditions is used to classify colors in the modified YUV color space for the robot soccer vision system. For this, a new method to measure brightness is proposed by use of a color card. After the neural network is constructed, a look-up-table is generated to replace the neural network in order to reduce the computation time. Experimental results show that the proposed color classification method is robust under illumination variations.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.7 no.3
• /
• pp.143-150
• /
• 2012

In this paper, we introduce an intelligent quadruped walking robot that can perform stable walking and a couple of intelligent behaviors. The developed robot has two sets of ultrasonic sensors and six sets of infrared sensors and can perform obstacle avoidance by detecting obstacles and estimating the distances and directions of those obstacles. The robot also has a stereo camera and can paly soccer by detecting a ball and estimating the 3 dimensional coordinates of the ball. In performing those intelligent behaviors, the robot needs to have the capability of generating its walking patterns, solving the inverse kinematics problem, and interfacing several sensors in realtime. Therefore we designed a hierarchical controller that consists of a main controller and an auxiliary controller. The main controller is a 32-bit DSP that can perform fast floating-point opertaion and the auxiliary one is a 8-bit micro-controller. We showed that the developed quadruped walking robot successfully perform those intelligent behaviors through experiments.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2001.04a
• /
• pp.177-181
• /
• 2001

This paper presents an strategy algorithm of a soccer robot. We simply classified strategy of soccer robot as attack and defense. We use DC-motor in our Soccer Robot. We use the vision system made by MIRO team of Kaist and Soty team for image processing. Host computer is made by Pentium III. The RF module is used for the communication between each robot and the host computer. Fuzzy logic is applied to the path planning of our robot. We improve strategy algorithm of soccer robot. Here we explain improvement of strategy algorithm and fault of the our soccer robot system.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.47-50
• /
• 2003

기계학습의 한 종류인 분류자 시스템은 간단한 문제에 대하여 실시간 처리와 온라인 학습이 가능하다. 그러나 복잡한 환경에서는 빠른 적응이 힘들다. 본 논문에서는 복잡한 환경에서 분류자 시스템의 적응 성능을 개선함으로써 실시간이 가능하도록 전체 환경을 분류하고 각기 다른 룰 셋을 이용하는 룰 클러스터링에 의한 분류자 시스템을 제안한다 환경을 상황에 따라 나눔으로써 전체 환경이 변화하였을 경우 각 상황에 따른 변화에 대해서만 추가적으로 학습함으로써 탐색 공간을 줄여 학습 시간을 감소시킨다. 제안한 시스템은 분류자 시스템 중 ZCS을 이용하여 로봇축구 시스템에 적용하여 기존의 방법과 그 성능을 비교 검토한다.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.24 no.6
• /
• pp.696-702
• /
• 2015

This paper suggests a new method for making a navigation path by using Bezier curve in order to improve the navigation performance used to avoid obstacles during a robot soccer game. We analyzed the advantages and disadvantages of both vector-field and limit-cycle navigation methods, which are the mostly widely used navigation methods for avoiding obstacles. To improve the disadvantages of these methods, we propose a new design technique for generating a more proper path using Bezier curve and describe its advantages. Using computer simulations and experiments, we compare the performance of vector-field navigation with that of Bezier curve navigation. The results prove that the navigation performance using Bezier curve is relatively superior to the other method.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.4 no.1
• /
• pp.9-15
• /
• 2009

In this paper, a network based robot soccer system is proposed. The system consists of robots, an image processing sub-system, a game server, and client systems. Embedded technique is applied to the hardware and software for controlling the robots and image processing. In this robot soccer system, a gamer can see and control robots in a remote site through Internet. During the game, the game server gives geometrical information on robots such as positions and orientations. We demonstrated the game in public and obtained optimistic results even though some technical problemssuch as communication delay and precise control for the robots should be improved.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.4 no.3
• /
• pp.118-125
• /
• 2009

This paper proposes a novel method to enhance performance of soccer robot system using optimal color patch mounted on the robot. In soccer robot system, the position and orientation of the robot can be estimated with color patch under real time environment. However, the location estimation of the robot is very sensitive to the pattern of color patch. In addition, pattern recognition and navigation algorithm are operated independently to reduce the operation time. The experimental results show that the proposed pattern of patch is effective to reduce the position and orientation error of the robot.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.106-108
• /
• 2006

강화 학습은 누적 보상 값을 최대화할 수 있는 행동 선택 전략을 학습하는 온라인 학습의 한 형태이다. 효과적인 강화학습을 위해 학습 에이전트가 매 순간 고민해야 하는 문제가 탐험(exploitation)과 탐색(exploration)의 문제이다. 경험과 학습이 충분치 않은 상태의 에이전트는 어느 정도의 보상 값을 보장하는 과거에 경험한 행동을 선택하느냐 아니면 보상 값을 예측할 수 없는 새로운 행동을 시도해봄으로써 학습의 폭을 넓힐 것이냐를 고민하게 된다. 특히 단일 에이전트에 비해 상태공간과 행동공간이 더욱 커지는 다중 에이전트 시스템의 경우, 효과적인 강화학습을 위해서는 상태 공간 축소방법과 더불어 탐색의 기회가 많은 행동 선택 전략이 마련되어야 한다. 본 논문에서는 로봇축구 Keepaway를 위한 효율적인 다중 에이전트 강화학습 방법을 설명한다. 이 방법의 특징은 상태 공간 축소를 위해 함수근사방법의 하나인 타일 코딩을 적용하였고, 다양한 행동 선택을 위해 룰렛 휠 선택 전략을 적용한 것이다. 본 논문에서는 이 방법의 효과를 입증하기 위한 실험결과를 소개한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.369-372
• /
• 2002

다양한 특성들을 가지고 있는 멀티에이전트 시스템의 행동학습은 에이전트 설계에 많은 부담을 덜어준다. 특성들로부터 나오는 다양한 행동의 효과적인 학습은 에이전트들이 환경에 대한 자율성과 반응성을 높여준 수 있다. 행동학습은 model-based learning과 같은 교사학습보다는 각 상태를 바로 지각하여 학습하는 강화학습과 같은 비교사 학습이 효과적이다. 본 논문은 로봇축구환경에 에이전트들의 행동을 개선된 강화학습법인 Modular Q-learning을 적용하여 복잡한 상태공간을 효과적으로 나누어 에이전트들의 자율성과 반응성을 높일 수 있는 강화학습구조를 제안한다.

• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.2998-3000
• /
• 1999

Robot soccer is a challenging research area in which multiple robots collaborate in adversarial environment to achieve specific objectives. We designed and built the robotic agents for robot soccer, especially MIROSOT. We have been developing the appropriate vision algorithm, algorithm for ball tracking and prediction, algorithms for collaboration between the robots in an uncertain dynamic environment. In this work we focus on the development of ball tracking and prediction algorithm using Kalman filter. Robustness and feasibility of the proposed algorithm is demonstrated by simulation.