http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
조백규(BaekKyu Cho) 제어로봇시스템학회 2012 제어·로봇·시스템학회 논문지 Vol.18 No.1
Human generally uses three methods to keep balance. One of them is using reactive momentum such as swing an upper body or arms. In this study, we proposed a balancing controller for the reactive momentum method using an inverted pendulum. We simplified a human or a humanoid robot as a two-link inverted pendulum having two edges. In addition, we proposed a distinctive condition for controller transition. If a human is pushed, he has to change a balancing controller from using an ankle torque to using a reactive momentum or changing foot placement. When the balancing controller is changed from using an ankle torque to using a reactive momentum, it is required a proper timing to keep a stability and make smooth movement. In the experiment, the proposed controller and distinctive condition were verified.
조백규(Baek Kyu Cho) 제어로봇시스템학회 2012 제어·로봇·시스템학회 논문지 Vol.18 No.9
We introduce a pattern generation method for a hopping one-legged robot and verify it experimentally. The pattern is derived from the liner and angular momentum of a COM (Center of Mass), which are pre-scheduled. Because of the relation between angular velocities of joints and momemtums of the COM, joint angle trajectories are easily obtained. In addition, the landing impact force is reduced by only adjusting the landing timing. In the experiment, the one-legged robot hops in place with 0.06 s of flying time, and makes continuous hopping. Based on our experimental results, the proposed method can be applied to hopping and running of biped humanoid robots.
전호진(Ho-Jin Jeon),조백규(Baek-Kyu Cho) 제어로봇시스템학회 2017 제어·로봇·시스템학회 논문지 Vol.23 No.5
Image processing is widely used not only in manufacturing industries, but also in advanced contexts such as RoboCup, an international robotics competition. Recently, in the image processing field, convolutional neural networks, which is a deep learning approach, became the mainstream of object recognition algorithms. In this paper, a convolutional neural network is designed and used to learn the features of the ball in the RoboCup soccer game. The convolutional neural network was named JeoNet. JeoNet was modified to learn fewer features than VGGNet and ResNet, but shows the same performance. In order to obtain the detected ball’s position, the Single Shot Multibox Detector was applied. To verify JeoNet, an experimental environment was constructed in which a soccer robot finds a ball in the sight. The benefits of JeoNet were shown by comparing the tracking time of the ball between JeoNet and conventional machine vision based ball finding algorithms.
가변 샘플링 시간 모델 예측 제어(MPC)를 기반으로 계단 등반 로봇 MOBINN
김건오(Geono Kim),조백규(Baek-Kyu Cho) 국방로봇학회 2024 국방로봇학회 논문집 Vol.3 No.2
This paper presents an innovative approach to enhance stair climbing capabilities in mobile robots, focusing on MOBINN, a two-wheeled balancing robot designed for both indoor and outdoor environments. Given the increasing demand for robots capable of navigating complex terrains, including stairs, we introduce a novel control strategy that employs a Variable Sampling Time Model Predictive Control (MPC) framework. This methodology optimizes the robots performance by adjusting the sampling time dynamically, based on the stair information anticipated by the robots sensors. Through precise balance control and efficient torque generation, MOBINN can adapt to various stair dimensions and conditions, overcoming the limitations of conventional wheeled robots. The paper details the development of the control framework, including the dynamics of stair climbing, formulation of the variable sampling time Model Predictive Control (MPC), and experimental validation. Results demonstrate MOBINNs enhanced stair climbing efficiency and stability, showcasing the potential of variable sampling time MPC in robotic mobility applications. This study contributes to the field by providing a scalable solution for robots navigating staircases, a critical challenge in urban and domestic environments.
풍력발전기 진단 및 청소를 위한 로봇의 모델링 및 자세제어
공진영(Jin-Young Kong),이재순(Jae-Soon Lee),강연식(Yeon-Sik Kang),조백규(Baek-Kyu Cho) 제어로봇시스템학회 2014 제어·로봇·시스템학회 논문지 Vol.20 No.9
Wind turbines are in the limelight in the alternative energy industry. However, they face frequent and various problems during operation. We focused on the supervising of the blades of a wind turbine. In this paper, we present the design of a maintenance robot that takes the size of wind turbine blades into consideration, so the general form of the robot is a square with four wires fixed to its vertices and to the nacelle. After the robot is placed near the nacelle, it moves along the blades. We also designed an attitude control algorithm for the robot to maintain its balance. Our control algorithm for the robot consists of roll and pitch attitude controllers and a height controller. Each controller was designed independently and then superposed together. We used simulations to verify our control algorithm.
직접 메탄올 연료전지를 이용한 휴머노이드 로봇의 전원시스템 개발
조한익(Joh, Han-Ik),하태정(Ha, Tae-Jung),조재형(Cho, Jae-Hyung),김종호(Kim, Jong-Ho),채승훈(Chae, Seung-Hoon),황상엽(Whang, Sang-Youp),문상흡(Moon, Sang-Heup),하흥용(Ha, Heung-Yong),조백규(Cho, Baek-Kyu),오준호(Oh, Jun-Ho) 한국신재생에너지학회 2006 한국신재생에너지학회 학술대회논문집 Vol.2006 No.11
전극 전해질 복합체가 하나로 구성된 단위전지는 전기출력이 낮아 원하는 전기출력을 얻기 위하여 단위전지를 여러 장 직렬로 쌓아 스택을 구성하여야 한다. 본 연구에서는 400W급 직접메탄을 연료전지 스택을 제조하여, 운전 조건의 최적화를 위한 실험을 수행하였고, 한국형 휴머노이드 로봇인 Hubo에 적용하여 그 동특성을 평가하였다. 0.8M 메탄을 용액은 260 ml/min의 유량을, 공기는 42LPM을 스택에 주입하여 400W의 성능을 나타내었다. 연료전지 스택의 온도와 전압의 편차는 작았으며 정상 운전시스템의 온도는 약 70?C를 유지하였다. 로봇의 필요 전력이 높을 경우, 배터리와 연료전지가 동시에 부하 변동에 대응하였다 방전된 배터리는 로봇의 필요 전력이 낮아질 경우, 충전되어 이후의 방전 상황에 대비하였다. 연료전지 시스템 적용 결과 로봇의 성공적인 운전을 확인할 수 있었다.