• The Journal of the Korea Contents Association
• /
• v.15 no.3
• /
• pp.57-67
• /
• 2015

The aim of this study was to categorize and analyze the patterns of the observation type in the experiential learning through the science museum for elementary gifted students in science. Ten science-gifted students were included and analyzed in this study and during experiential learning in the science museum, the participants freely expressed their observation of their own languages and all observations and dialogue were recorded. The results are listed below. The cognitive aspect, especially question and response activity without their personal opinion, was the most frequently used item. Among the affective aspects, item for 'recommendation' was often used. In accordance to observation type, most participants overall observed single object independently of time. Also, participants mostly observed objects visually using qualitative method without manipulation. Therefore verbal interaction through question might have a positive effect on frequency and diversity of observation. Project learning, such as particular exhibition hall, exploratory time of concentration by students, or study paper will be capable of creating a effective observation learning in order to induce a variety of observation of science gifted students in the experiential learning through the science museum.

• Journal of Power System Engineering
• /
• v.16 no.5
• /
• pp.70-75
• /
• 2012

In mechanical industries, forging is one of the basic process. But comparing the other developed industries, forging industries can not reach at the level of that development. In forging industries, the quality of the products totally depends on the skills of workers and also the precision of the equipments. Particularly because the open die forging industry is unable to deviate from the past method of production and all works are manually progressed, the operator is always exposed to the danger. In the regard some additional device has been made especially. Thus, in this research, by using the forklift as the means for the manipulation of the development object system, it tries to be comprised the process automation. After than it is fitted with the forklift for safe and easy handling of jobs and products during open die forging process. First of all, development system mold has been assembled to the system, after than it is assembled with forklift. This development system has been applied for handling of large scale products more than 300kg, and the satisfactory result with uniform quality of the products have been achieved due to this mechanical setup.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.854-857
• /
• 2020

최근 사람형(Anthropomorphic)로봇손의 사물조작 지능을 개발하기 위하여 행동복제(Behavior Cloning) Deep Reinforcement Learning(DRL) 연구가 진행중이다. 자유도(Degree of Freedom, DOF)가 높은 사람형 로봇손의 학습 문제점을 개선하기 위하여, 행동 복제를 통한 Human Demonstration Augmented(DA)강화 학습을 통하여 사람처럼 사물을 조작하는 지능을 학습시킬 수 있다. 그러나 사물 조작에 있어, 의미 있는 파지를 위해서는 사물의 특정 부위를 인식하고 파지하는 방법이 필수적이다. 본 연구에서는 딥러닝 YOLO기술을 적용하여 사물의 특정 부위를 인식하고, DA-DRL을 적용하여, 사물의 특정 부분을 파지하는 딥러닝 학습 기술을 제안하고, 2 종 사물(망치 및 칼)의 손잡이 부분을 인식하고 파지하여 검증한다. 본 연구에서 제안하는 학습방법은 사람과 상호작용하거나 도구를 용도에 맞게 사용해야하는 분야에서 유용할 것이다.

• The Journal of Society for e-Business Studies
• /
• v.18 no.4
• /
• pp.45-65
• /
• 2013

A design file can get larger in size as the complexity of the target object increases. A large design file may reside in a large parallel computing system, such as cloud computing systems, and many designers may work concurrently on the same design file. In such a case, it is obvious that we need some kind of protection mechanism so that each user can access only the area of the file he or she is entitled to. Two approaches can be taken for this problem: one is the traditional access control mechanisms and the other encryption techniques. We take the latter approach to ensure the safety of the file even in public domain such as clouding systems, and in this paper, we suggest an encryption scheme for a file where the file is encrypted in multi-layer so that each user is allowed to access the file only at the layer for which the user has the proper access right. Each layer of the file is encrypted with different keys and these keys are exposed only to those who have the right access permit. The paper explains the necessary file format to achieve this goal and discusses the file manipulation functions to handle this new file format.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.2
• /
• pp.508-516
• /
• 2010

An autonomous and automatic home mess-cleanup robot is newly developed in this paper. Thus far, vacuum-cleaners have lightened the burden of household chores but the operational labor that vacuum-cleaners entail has been very severe. Recently, a cleaning robot was commercialized to solve but it also was not successful because it still had the problem of mess-cleanup, which pertained to the clean-up of large trash and the arrangement of newspapers, clothes, etc. Hence, we develop a new home mess-cleanup robot (McBot) to completely overcome this problem. The robot needs the capability for agile navigation and a novel manipulation system for mess-cleanup. The autonomous navigational system has to be controlled for the full scanning of the living room and for the precise tracking of the desired path. It must be also be able to recognize the absolute position and orientation of itself and to distinguish the messed object that is to be cleaned up from obstacles that should merely be avoided. The manipulator, which is not needed in a vacuum-cleaning robot, has the functions of distinguishing the large trash that is to be cleaned from the messed objects that are to be arranged. It needs to use its discretion with regard to the form of the messed objects and to properly carry these objects to the destination. In particular, in this paper, we describe our approach for achieving accurate localization using RFID for home mess-cleanup robots. Finally, the effectiveness of the developed McBot is confirmed through live tests of the mess-cleanup task.

• Therapeutic Science for Rehabilitation
• /
• v.5 no.2
• /
• pp.23-33
• /
• 2016

Objectives: The aim of this study was to investigate the association between cognitive function and upper extremity function in the elderly. The articles were analyzed based on patient, intervention, comparison, and outcome using the P.I.C.O. principle. Methods: We systematically examined papers from January, 2000 to November, 2015 published papers through the foreign journals which were Medline & Pubmed for three months. mainly used key words were elderly, dementia, Alzheimer's disease, Mild cognitive impairment, age-related, aging, cognitive, upper extremity function, hand function, hand-grip strength, grip force, complex motor function, bimanual, dexterity, UE performance, and coordination. Results: The number of discovered outcomes for association between cognitive function and upper extremity function in the elderly was 7; grip strength & sex are impact on manipulation object, 1. The results show that cognitive function is associated with upper extremity function in the elderly. Conclusion: This study is expected to help selecting intervention, assessment tools according to the individual's degree of cognitive level and upper extremity function. In future domestic research, variety assessment tools need to be used and more qualitative level experiment will be carried out.

• Therapeutic Science for Rehabilitation
• /
• v.8 no.1
• /
• pp.51-62
• /
• 2019

Objective: To develop and verify the usability of a cognitive rehabilitation system with diverse cognitive functional levels based on tangible objects for the elderly population. Methods: A study was conducted to investigate the system's strengths and weaknesses by upgrading it with responses from two groups of 15 patients and 4 occupational therapists. After undergoing three forms of training - regarding executive function, memory, and concentration for a total of 20-30 min, the participants were asked to answer a structured questionnaire about contents of the three forms of training, hardware including the tablet PC functioning as a CPU and display media and tangible objects, and satisfaction of experiential usage of the system. Results: Both groups responded that the most interesting training area was executive function while the least interesting was concentration. Six participants reported that the size of the screen of the tablet PC was inappropriate, and five responded that the size of the tool was inappropriate. All therapists and 40% of the patients responded that they were satisfied with this system. Conclusion: This system's features include easy manipulation of tangible tools for performing training tasks, easy selection of and training in cognitive areas based on users' needs, and automatic adjustment of difficulty level based on users' performance. The training environment was designed to be similar to the natural environment by using tangible objects in both hands as input devices for the system, and the system was considered as an alternative to the lack of community cognitive rehabilitation specialists.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.3
• /
• pp.141-148
• /
• 2022

Recently, reinforcement learning combined with deep neural network technology has achieved remarkable success in various fields such as board games such as Go and chess, computer games such as Atari and StartCraft, and robot object manipulation tasks. However, such deep reinforcement learning describes states, actions, and policies in vector representation. Therefore, the existing deep reinforcement learning has some limitations in generality and interpretability of the learned policy, and it is difficult to effectively incorporate domain knowledge into policy learning. On the other hand, dNL-RRL, a new relational reinforcement learning framework proposed to solve these problems, uses a kind of vector representation for sensor input data and lower-level motion control as in the existing deep reinforcement learning. However, for states, actions, and learned policies, It uses a relational representation with logic predicates and rules. In this paper, we present dNL-RRL-based policy learning for transportation mobile robots in a manufacturing environment. In particular, this study proposes a effective method to utilize the prior domain knowledge of human experts to improve the efficiency of relational reinforcement learning. Through various experiments, we demonstrate the performance improvement of the relational reinforcement learning by using domain knowledge as proposed in this paper.

• Korean Journal of Cognitive Science
• /
• v.33 no.4
• /
• pp.169-203
• /
• 2022

As artificial intelligence (AI) technology advances, the number of cases in which AI becomes an object or subject of moral judgment is increasing, and this trend is expected to accelerate. Although the area of AI in human society expands, relatively few studies have been conducted on how people perceive and respond to AI. Three studies examined the effect of the anthropomorphism of AI on its responsibility. We predicted that anthropomorphism would increase the responsibility perception, and perceived agency and perceived patiency for AI would mediate this effect. Although the manipulation was not effective, multiple analyses confirmed the indirect effect of perceived patiency. In contrast, the effect of perceived agency of AI was somewhat mixed, which makes the hypothesis partially supported by the overall result. This result shows that for the moral status of artificial agents, perceived patiency is relatively more critical than perceived agency. These results support the organic perspective on the moral status that argues the importance of patiency, and show that patiency is more important than agency in the anthropomorphism related study of AI and robots.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.858-861
• /
• 2020

로봇이 사람과 같이 다양하고 복잡한 사물 조작을 하기 위해서 휴먼형 로봇손의 사물 파지 작업이 필수적이다. 자유도 (Degree of Freedom, DoF)가 높은 휴먼형(anthropomorphic) 로봇손을 학습시키기 위하여 사람 데모(human demonstration)가 결합된 강화학습 최적화 방법이 제안되었다. 본 연구에서는 강화학습 최적화 방법에 사람 데모가 결합된 Demonstration Augmented Natural Policy Gradient(DA-NPG)와 NPG 의 성능 비교를 통하여 행동 복제의 효율성을 확인하고, DA-NPG, DA-Trust Region Policy Optimization (DA-TRPO), DA-Proximal Policy Optimization (DA-PPO)의 최적화 방법의 성능 평가를 위하여 6 종의 물체에 대한 휴먼형 로봇손의 사물 조작 작업을 수행한다. 그 결과, DA-NPG 와 NPG를 비교한 결과를 통해 휴먼형 로봇손의 사물 조작 강화학습에 행동 복제가 효율적임을 증명하였다. 또한, DA-NPG 는 DA-TRPO 와 유사한 성능을 보이면서 모든 물체에 대한 사물 파지에 성공하여 가장 안정적이었다. 반면, DA-TRPO 와 DA-PPO 는 사물 조작에 실패한 물체가 존재하여 불안정한 성능을 보였다. 본 연구에서 제안하는 방법은 향후 실제 휴먼형 로봇에 적용하여 휴먼형 로봇 손의 사물조작 지능 개발에 유용할 것으로 전망된다.