• The Journal of Korea Robotics Society
• /
• v.15 no.4
• /
• pp.404-412
• /
• 2020

Typically everyday human life tasks involve at least two people moving objects such as tables and beds, and the balancing of such object changes based on one person's action. However, many studies in previous work performed their tasks solely on robots without factoring human cooperation. Therefore, in this paper, we propose cooperative robot for table balancing using Q-learning that enables cooperative work between human and robot. The human's action is recognized in order to balance the table by the proposed robot whose camera takes the image of the table's state, and it performs the table-balancing action according to the recognized human action without high performance equipment. The classification of human action uses a deep learning technology, specifically AlexNet, and has an accuracy of 96.9% over 10-fold cross-validation. The experiment of Q-learning was carried out over 2,000 episodes with 200 trials. The overall results of the proposed Q-learning show that the Q function stably converged at this number of episodes. This stable convergence determined Q-learning policies for the robot actions. Video of the robotic cooperation with human over the table balancing task using the proposed Q-Learning can be found at http://ibot.knu.ac.kr/videocooperation.html.

• Asian Journal of Atmospheric Environment
• /
• v.12 no.2
• /
• pp.178-191
• /
• 2018

The concentrations and chemical compositions of volatile organic compounds (VOCs), including volatile fatty acids, phenols, indoles, aldehydes, and ketones, which are the main organic compounds generated by swine, were investigated in July and October 2016 and January 2017. In addition, the emission rates and annual emissions of these components from the swine shed were estimated. The concentrations of VOCs in the swine shed averaged $511.3{\mu}g\;m^{-3}$ in summer, $315.5{\mu}g\;m^{-3}$ in fall and $218.6{\mu}g\;m^{-3}$ in winter. Acetone, acetic acid, propionic acid, and butyric acid were the predominant components of the VOCs, accounting for 80-88% of the total VOCs. The hourly variations of VOC concentrations in the swine shed in fall and winter suggest that the VOC concentrations were related to the ventilation rate of the swine shed, the activity of the swine, and the temperature in the swine shed. Accordingly, the emission rates of VOCs from the swine shed were $1-2{\times}10^3{\mu}g(h\;kg-swine)^{-1}$.

• Proceedings of the Korean Society of Tribologists and Lubrication Engineers Conference
• /
• 2001.06a
• /
• pp.271-280
• /
• 2001

An NAO friction material (low-steel type) containing 15 ingredients was investigated to study the role of transfer film on the friction characteristics. The friction material specimens with extra 100% of each ingredient were tested using a pad-on-disk type tribotester. A non-destructive method of measuring the transfer film was developed by considering the electric resistance of the transfer film. Results showed that solid lubricants and iron powder assisted transfer film formation on the rotor surface. Average friction coefficient was independent of transfer film thickness in this experiment. On the other hand, the thick transfer film on the rotor surface reduced the amplitude of friction oscillation under temperature conditions ( 250$^{\circ}C$) that transfer film forms.

• Proceedings of the Korean Society of Tribologists and Lubrication Engineers Conference
• /
• 2001.06a
• /
• pp.72-78
• /
• 2001

An experimental investigation was conducted to examine the tribological behavior of NAO (non-asbestos organic) type brake linings containing different volume ratios of graphite and antimony trisulfide (Sb$_2$S$_3$). In order to investigate the effect of the solid lubricants on brake performance, three different friction tests (pressure, speed, and temperature sensitive tests) were carried out using a scale dynamometer. The test results showed that the friction characteristics were strongly affected by the type and the amount of solid lubricants in the brake lining. It was found that the brake linings with both solid lubricants were better in friction stability due to the complementary role of the two disparate lubricating properties at various pressure and speed conditions. In particular, the brake lining containing higher concentrations of graphite showed better fade resistance than others during high temperature friction test.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.14-14
• /
• 2018

세계의 여러 국가에서 과거 발생했던 강수의 통계적 특성에서 벗어나는 극치사상이 빈번하게 관측되고 있다. 이와 같은 현상에 가장 큰 영향을 미치고 있는 요인중 하나는 지구온난화이며 실제 산업화 이후 온실가스의 증가와 더불어 극한 기상현상의 발생 빈도가 증가하였다. 현재 예상치 못한 수문사상의 발생으로 인해 수자원관리에 있어서 많은 어려움을 겪고 있으며, 특히 호우사상은 막대한 인명 및 사회적 피해를 야기하고 있다. 우리나라의 경우 계절적 특징으로 여름철에 강수가 집중되는 양상을 보이고 있으며 따라서 여름철 강수량을 예측하여 호우에 대한 대비책을 마련해야한다. 계절강수 예측은 수문, 산림, 식품, 등을 포함한 사회 경제적 파급 효과가 매우 크지만 아직 신뢰성 있는 예측은 어려운 상태이다. 또한, 발생 강도와 빈도가 큰 극한 강우는 주로 짧은 시간에 걸쳐 발생하기 때문에 예측하기가 어렵다. 최근 다양한 분야의 연구에서 AO, NAO, ENSO, PDO등과 같은 외부적 요인이 수문학적 빈도를 변화시킨다고 알려지고 있어 본 연구에서는 Bayesian 통계기법을 이용한 비정상성 빈도해석모형을 토대로 외부 기상인자에 의한 변동성을 고려할 수 있는 계절강수량 예측모형을 구축한 후 산정된 결과를 입력 자료로 하여 극치강수량을 추정할 수 있는 비정상성 Four - Parameter (4P)-Beta분포를 이용한 알고리즘을 개발하여 직접적으로 일단위 이하의 극치강수량을 상세화 시킬 수 있는 모형으로 확장하여 이를 통해 기상변동성을 다양한 시간규모에서 고려하기 위한 정보로 활용하고자 하였다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.211-211
• /
• 2018

미해양대기청 기후예측센터(Climate Prediction Center, NOAA)에서 제공하고 있는 기후지수(climate indices)를 예측인자로 하고 금강유역의 5~6월의 강우량을 예측대상으로 하는 원격상관기반 통계모형을 구축하였다. 1988년부터 2017년까지의 30년 자료에 대해 예측인자와 예측대상간의 시간지연상관분석을 수행한 결과 NAO(North Atlantic Oscillation), EP/NP(East Pacific/North Pacific Oscillation), EA(East Atlantic Pattern), WP(Western Pacific Index) 등과 상관성이 높은 것으로 분석되었으며, 이러한 시간지연 기후지수를 이용하여 4개월전에 5,6월 강수량을 예측할 수 있는 다중회귀모형을 개발하였다. 관측 강우량 아노말리가 큰 경우에는 다소 과소 예측되고, 아노말리가 작은 경우에는 다소 과다 예측되는 경향을 보였지만 관측 강우량과 예측 강우량간의 상관계수가 0.75로서 비교적 우수한 예측 결과를 나타내었다. 5~6월 강우량 아노말리의 3분위 예측성을 평가한 결과 평년이상 적중률은 77.8%, 평년수준은 81.8%로서 예측 성공률이 높았으며, 5, 6월 누적강우량이 매우 작았던 92년과 95년을 제외하고는 강우량이 적은 해에도 예측성이 우수하여 평년이하 적중률이 70.0%를 나타내었다. 따라서 본 개발모형은 최소 4개월 이전 선행시간을 가지고 늦봄, 초여름강우량을 예측할 수 있는 저비용의 가뭄 예측 도구로 유용하게 활용될 수 있을 것이다.

• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.264-272
• /
• 2022

Human-robot cooperative tasks are increasingly required in our daily life with the development of robotics and artificial intelligence technology. Interactive reinforcement learning strategies suggest that robots learn task by receiving feedback from an experienced human trainer during a training process. However, most of the previous studies on Interactive reinforcement learning have required an extra feedback input device such as a mouse or keyboard in addition to robot itself, and the scenario where a robot can interactively learn a task with human have been also limited to virtual environment. To solve these limitations, this paper studies training strategies of robot that learn table balancing tasks interactively using deep reinforcement learning with human's facial expression feedback. In the proposed system, the robot learns a cooperative table balancing task using Deep Q-Network (DQN), which is a deep reinforcement learning technique, with human facial emotion expression feedback. As a result of the experiment, the proposed system achieved a high optimal policy convergence rate of up to 83.3% in training and successful assumption rate of up to 91.6% in testing, showing improved performance compared to the model without human facial expression feedback.

• BMB Reports
• /
• v.53 no.5
• /
• pp.260-265
• /
• 2020

Scorpion venom comprises a cocktail of toxins that have proven to be useful molecular tools for studying the pharmacological properties of membrane ion channels. HelaTx1, a short peptide neurotoxin isolated recently from the venom of the scorpion Heterometrus laoticus, is a 25 amino acid peptide with two disulfide bonds that shares low sequence homology with other scorpion toxins. HelaTx1 effectively decreases the amplitude of the K⁺ currents of voltage-gated Kv1.1 and Kv1.6 channels expressed in Xenopus oocytes, and was identified as the first toxin member of the κ-KTx5 subfamily, based on a sequence comparison and phylogenetic analysis. In the present study, we report the NMR solution structure of HelaTx1, and the major interaction points for its binding to voltage-gated Kv1.1 channels. The NMR results indicate that HelaTx1 adopts a helix-loop-helix fold linked by two disulfide bonds without any β-sheets, resembling the molecular folding of other cysteine-stabilized helix-loop-helix (Cs α/α) scorpion toxins such as κ-hefutoxin, HeTx, and OmTx, as well as conotoxin pl14a. A series of alanine-scanning analogs revealed a broad surface on the toxin molecule largely comprising positively-charged residues that is crucial for interaction with voltage-gated Kv1.1 channels. Interestingly, the functional dyad, a key molecular determinant for activity against voltage-gated potassium channels in other toxins, is not present in HelaTx1.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.01a
• /
• pp.145-150
• /
• 2001

A Controlled Ecological Life Support System(CELSS) is essential for man to live a long time in a closed space such as a lunar base or a mars base. Such a system may be an extremely complex system that has a lot of facilities and circulates multiple substances,. Therefore, it is very difficult task to control the whole CELSS. Thus by regarding facilities constituting the CELSS as agents and regarding the status and action as information, the whole CELSS can be treated as multi-agent system(MAS). If a CELSS can be regarded as MAS the CELSS can have three advantages with the MAS. First the MAS need not have a central computer. Second the expendability of the CELSS increases. Third, its fault tolerance rises. However it is difficult to describe the cooperation protocol among agents for MAS. Therefore in this study we propose to apply reinforcement learning (RL), because RL enables and agent to acquire a control rule automatically. To prove that MAS and RL are effective methods. we have created the system in Java, which easily gives a distributed environment that is the characteristics feature of an agent. In this paper, we report the simulation results for material circulation control of the CELSS by the MAS and RL.

• Journal of Astronomy and Space Sciences
• /
• v.26 no.4
• /
• pp.511-520
• /
• 2009

In this study, the Ocean Tide Loading (OTL) constituents were detected by the Precise Point Positioning (PPP) technique using GPS. Then, the GPS estimates of OTL constituents were compared with the predictions of the ocean tide models. We picked three permanent GPS stations as test sites and they are ICNW, SEOS, and CJUN. To detect the OTL constituents using GPS, we created vertical coordinate time series at 10-minute intervals using the PPP approach implemented in the GIPSY software. Through the tidal harmonic analysis of this height time series, the four major constituents ($M_2$, $S_2$, $K_1$, $O_1$) were determined. The amplitude obtained from the GPS height time series of the OTL constituents showed best match with the model predictions at CJUN, while the phase showed closest match at ICNW. The amplitude accuracy of the $M_2$, which is the dominant factor out of the 11 major constituents, was 24.8% on average.