• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.759-762
• /
• 2018

최근 들어 점차 지능형 서비스 로봇들이 인간의 실생활 속으로 들어옴에 따라, 로봇 스스로 다양한 물체들을 효과적으로 조작할 수 있는 지식을 습득하는 기계 학습 기술들이 매우 주목을 받고 있다. 전통적으로 로봇 행위 학습 분야에는 강화 학습 혹은 심층 강화 학습 기술들이 주로 많이 적용되어 왔으나, 이들은 대부분 물체 조작 작업과 같이 다차원 연속 상태 공간과 행동 공간에서 최적의 행동 정책을 학습하는데 여러가지 한계점을 가지고 있다. 따라서 본 논문에서는 전문가의 데모 데이터를 활용해 보다 효율적으로 물체 조작 행위들을 학습할 수 있는 모방 학습과 강화 학습의 통합 프레임워크를 제안한다. 이 통합 프레임워크는 학습의 효율성을 향상시키기 위해, 기존의 GAIL 학습 체계를 토대로 PPO 기반 강화 학습 단계의 도입, 보상 함수의 확장, 상태 유사도 기반 데모 선택 전략의 채용 등을 새롭게 시도한 것이다. 다양한 성능 비교 실험들을 통해, 본 논문에서 제안한 통합 학습 프레임워크인 PGAIL의 우수성을 확인할 수 있었다.

• The Journal of Korea Robotics Society
• /
• v.14 no.3
• /
• pp.163-169
• /
• 2019

Robot ALICE@ERICA is a service robot developed to receive donations and to provide information services. ALICE@ERICA stands for Artificial Learning Intelligence robot for Culture and Entertainment at ERICA. In order to achieve the specific purpose of receiving donations, proper appearance design, appropriate movement and good communication skills are required in terms of HRI. In this paper, we introduce three strategies for developing robots to receive donations effectively. The first is to design a robot that makes people feel intimacy, the second is to approach only one of several people as a donor, and finally the donor communicates with video contents and voice recognition. A survey was conducted on the person who showed the reaction after the robot donated money in public places. Based on the survey results, it is proved that the method presented in this study effectively contributed to fund raising. If robots can perform actions that require high level of HRI, such as donation, robots can contribute more to human society. We hope that this study contributes to the improvement of human happiness.

• Journal of Sensor Science and Technology
• /
• v.33 no.1
• /
• pp.18-23
• /
• 2024

Humanoid robots, designed to interact in human environments, require stable mobility to ensure safety. When a humanoid robot falls, it causes damage, breakdown, and potential harm to the robot. Therefore, fall detection is critical to preventing the robot from falling. Prevention of falling of a humanoid robot requires an operator controlling a crane. For efficient and safe walking control experiments, a system that can replace a crane operator is needed. To replace such a crane operator, it is essential to detect the falling conditions of humanoid robots. In this study, we propose falling detection methods using Convolution Neural Network (CNN) model. The image data of a humanoid robot are collected from various angles and environments. A large amount of data is collected by dividing video data into frames per second, and data augmentation techniques are used. The effectiveness of the proposed CNN model is verified by the experiments with the humanoid robot MAX-E1.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.7
• /
• pp.1418-1423
• /
• 2004

A legged robot is friendly to human because it is resemble to human. And the robot can obtain support points freely because it has high degree of freedom for several joint as compared with a wheeled robot. Also the robot can create the relative position at desired position between support position and robot. The joint of robot cu used as manipulator. On the contrary the mechanism of robot is complicated to have many joint and moving speed is lower than wheeled robot. Also the legged robot is needed a special control not to fall on the ground because the robot is easy to vibrate when it is moving. The 4 leg structure is the minimum leg numbers not to fall and to realize safety gait continuously. A trot gait is investigated through experiments using a quadruped walking robot named TITAN-VIII.

• 한국정보교육학회:학술대회논문집
• /
• 2010.08a
• /
• pp.15-18
• /
• 2010

As robots are proved to be effective in enhancing students' creativity and problem-solving abilities and satisfying various needs in special education for the gifted, many students participate in private education and after-school robot classes. However, it is difficult for students in the lower social economy class to use robots for their learning because of the high expense of robots. On this point, as a part of u-Learnng project, this research attempts to provide students in the lower social economy class with the opportunities to use robots for one year. At the end of the year, we will compare the experimental group and the control group in order to examine learning effects of using robots. Until now we have found many cases that show positive effects of the use of robots in students' learning.

• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.376-384
• /
• 2017

In this paper, we describe a case of using robot technology in child studies to examine children's judgement and reasoning of the life phenomenon on boundary objects. In order to control the effects of the appearance of the robot, which children observe or interact directly with, on the children's judgement and reasoning of the life phenomenon, we developed a robot similar to human. Unit experimental scenarios representing biological and psychological properties were implemented based on control of robot's motion, speech, and facial expressions. Experimenters could combine these multiple unit scenarios in a cascade to implement various scenarios of the human-robot interaction. Considering that the experimenters are researchers of child studies, there was a need to develop a remote operation console that can be easily used by non-experts in the robot field. Using the developed robot platform, researchers of child studies could implement various scenarios by manipulating the biological and psychological properties of the robot based on their research hypothesis. As a result, we could clearly see the effects of robot's properties on children's understanding about boundary object like robots.

• Journal of Korea Multimedia Society
• /
• v.13 no.5
• /
• pp.651-662
• /
• 2010

This paper describes a method to identify objects for autonomous navigation of an outdoor mobile robot. To identify objects, the robot recognizes the object from an image taken by moving robot on outdoor environment. As a beginning, this paper presents the candidates for a segment of region to building of artificial object, sky and trees of natural objects. Then we define their characteristics individually. In the process, we segment the regions of the objects included by preprocessing using multiple features. Multiple features are HSI, line segments, context information, hue co-occurrence matrix, principal components and vanishing point. An analysis of building identifies the geometrical properties of building facet such as wall region, windows and entrance. The building as intersection in vertical and horizontal line segment of vanishing point extracts the mesh. The wall region of building detect by merging the mesh of the neighbor parallelograms that have similar colors. The property estimates the number of story and rooms in the same floors by merging skewed parallelograms of the same color. We accomplish the result of image segmentation using multiple features and the geometrical properties analysis of object through experiments.

• Journal of Broadcast Engineering
• /
• v.25 no.4
• /
• pp.537-544
• /
• 2020

This paper proposes the creation of an android robot head based on the facial action coding system(FACS), and the generation of emotional expressions by FACS. The term android robot refers to robots with human-like appearance. These robots have artificial skin and muscles. To make the expression of emotions, the location and number of artificial muscles had to be determined. Therefore, it was necessary to anatomically analyze the motions of the human face by FACS. In FACS, expressions are composed of action units(AUs), which work as the basis of determining the location and number of artificial muscles in the robots. The android head developed in this study had servo motors and wires, which corresponded to 30 artificial muscles. Moreover, the android head was equipped with artificial skin in order to make the facial expressions. Spherical joints and springs were used to develop micro-eyeball structures, and the arrangement of the 30 servo motors was based on the efficient design of wire routing. The developed android head had 30-DOFs and could express 13 basic emotions. The recognition rate of these basic emotional expressions was evaluated at an exhibition by spectators.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2017.04a
• /
• pp.143-143
• /
• 2017

산업이 고도화됨에 따라 자동화 기계 및 로봇에 의해 대량 생산 되는 품목과 달리, 작업 절차의 비정형성, 비연속성 등으로 인해 여전히 농업에 많은 인력이 투입되고 있다. 국제노동기구에 따르면, 세계 인력의 절반이 농업 인력에 해당하고 작업 중 부상이나 사망 등으로 인해 가장 위험한 직업군 중 하나에 해당하는 것으로 나타났다. 시설 재배 농업의 경우, 노동집약적인 온실 내 작업 특성상 잘못된 자세로 작업하거나 지나친 작업량 등으로 인해 작업자에게 근골격계 질환이 발생할 수 있다. 근골격계 질환으로 인해 작업효율이 감소하거나 생산비용의 증가로 이어질 수 있으며, 농가 수익에 손실이 발생할 수 있다. 이에 본 연구에서는 현행 시설 재배 농업에서 사용되는 레일 전동 작업차를 이용하여 작업자가 토마토를 수확할 때의 신체에 대한 농작업의 부하를 평가하고자 하였다. 작업차를 이용한 주요 작업 절차는 작물로부터 과실 수확, 과실 상자에 과실 투입, 빈 과실 상자와 가득 찬 과실 상자의 교대, 작업차 위의 과실 상자를 운반용 파레트에 하역하는 순서로 이루어지는 것을 확인하였다. 비디오장비로 촬영된 일련의 농작업 과정을 OWAS, RULA, REBA와 같은 체크리스트형 인간공학적 작업 부하 평가 도구를 이용하여 평가한 결과, 기존 레일 전동 작업차를 이용한 농작업의 근골격계 질환 유발 가능성을 확인하였다. 동작별 위험성을 토대로 근골격계 질환 유발 가능성이 높아 개선이 필요한 농작업 동작을 선정하였다. 선정된 동작은 실험실 내 환경에서 피실험자를 통한 모의 동작의 생체 신호 계측을 통해 신체 부하 정도를 정량적으로 측정할 수 있으며, 보조가 필요한 신체 부위를 특정하거나 안전성 확보가 필요한 동작에 대한 증거가 될 수 있다. 본 연구를 통해 향후 토마토 온실 내 신선도 유지를 위한 레일 전동 작업차의 개발에 작업자의 안전과 효율성 향상을 위한 인간공학적 설계를 적용할 수 있을 것으로 기대한다.

• Journal of the Korean Society for Precision Engineering
• /
• v.24 no.7 s.196
• /
• pp.90-97
• /
• 2007

This paper describes the development of 6-axis force/moment sensor considered adult weight far an intelligent foot of humanoid robot. In order to walk on uneven terrain safely, the foot should perceive the applied forces Fx, Fy, Fz and moments Mx, My, Mz to itself and control the foot using the forces and moments. The applied forces and moments should be measured from a 6-axis force/moment sensor attached to the foot, which is composed of Fx sensor, Fy sensor, Fz sensor, Mx sensor, My sensor and Mz sensor in a body. Each sensor should get the deferent rated load, because the applied forces and moments to foot in walking are deferent. Therefore, one of the important things in the sensor is to design each sensor with the deferent rated load and the same rated output. In this paper, a 6-axis force/moment sensor (rated load of Fx and Fy are 500Nm and Fz sensor is 1000N, and those of Mx and My are 18Nm, Mz sensor is 8Nm) for perceiving forces and moments in a humanoid robot's foot was developed using many PPBs (parallel plate-beams). The structure of the sensor was newly modeled, and the sensing elements (plate-beams) of the sensor were designed using by ANSYS software (FEM (Finite Element Method) program). Then, a 6-axis force/moment sensor was fabricated by attaching strain-gages on the sensing elements, and the characteristic test of the developed sensor was carried out. The rated outputs from FEM analysis agree well with that from the characteristic test.