• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.6
• /
• pp.518-523
• /
• 2015

This paper proposes a gait phase classifier using a Recurrent Neural Network (RNN). Walking is a type of dynamic system, and as such it seems that the classifier made by using a general feed forward neural network structure is not appropriate. It is known that an RNN is suitable to model a dynamic system. Because the proposed RNN is simple, we use a back propagation algorithm to train the weights of the network. The input data of the RNN is the lower body's joint angles and angular velocities which are acquired by using the lower limb exoskeleton robot, ROBIN-H1. The classifier categorizes a gait cycle as two phases, swing and stance. In the experiment for performance verification, we compared the proposed method and general feed forward neural network based method and showed that the proposed method is superior.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2000.04b
• /
• pp.365-370
• /
• 2000

A sliding mode fuzzy control (SMFC) algorithm is applied to design a controller for a benchmark problem on a wind- excited building. The structure is a 76-story concrete office tower with a height of 306 meters, hence the wind resistance characteristics are very important for the serviceability as well as the safety. A control system with an active tuned mass damper is assumed to be installed on the top floor. Since the structural acceleration is measured only at ,limited number of locations without measurement of the wind force, the structure of the conventional continuous sliding mode control may have the feed-back loop only. So, an adaptive least mean squares (LMS) filter is employed in the SMFC algorithm to generate a fictitious feed-forward loop. The adaptive LMS filter is designed based on the information of the stochastic characteristics of the wind velocity along the structure. A numerical study is carried out. and the performance of the present SMFC with the ,adaptive LMS filter is investigated in comparison with those of' other control, of algorithms such as linear quadratic Gaussian control, frequency domain optimal control, quadratic stability control, continuous sliding mode control, and H/sub ∞///sub μ/, control, which were reported by other researchers. The effectiveness of the adaptive LMS filter is also examined. The results indicate that the present algorithm is very efficient .

• Journal of Animal Environmental Science
• /
• v.4 no.1
• /
• pp.1-8
• /
• 1998

The objective of this study was conducted to develope a mobile pig nursery(MPN) for segregation of early weaned piglet. The units of MPN was consisted of 4 rectangle-type pens, 1 workroom, 2 doors, 3 windows, 1 air-inlet, 1 exhaust fan, and 1 cooling and heating system. Total of 40 piglet were weaned at 7 days of age. The period of the feeding test was 63 days. The results obtained are as follows: 1. The insulation thickness was 70, 70 and 50 mm of roof, floor and wall, respectively. R-value was 15.32 and 10.32 of roof and wall, respectively. 2. Exhaust fan( 30.48) was installed near back door 40cm from the bottom. When exhaust fan speeds were 15, 20, 30 and 35%, Air ventilation was required 9.75, 7.07, 1.72, 1.45 minutes of respectively. 3. Average temperature in the MPN was able to maintain 27∼28$^{\circ}C$ from 7∼28 days of age, 24∼25$^{\circ}C$ from 35∼56 days of age and 20∼21.5$^{\circ}C$ from 56∼70 days of age. 4. Average daily gain, feed intake and feed conversion were 420.6g, 761.5g, 1.81 respectively.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.1223-1226
• /
• 2006

P2P(Peer to Peer) 기술은 1990년대 후반기부터 산업계 및 학계에 주목을 받고 있는 기술 분야중의 하나로 이 기술의 장점은 인터넷 환경에 산재하여 있는 컴퓨팅 파워, 공간, 네트워크 대역을 인터넷 기반으로 효과적으로 활용하여 협력작업을 가능하게 한다는데 있다. 최근에는 모바일 환경 응용을 위한 P2P 디바이스 탐색 분야에 관심사가 증대되고 있으며, P2P 시스템은 중앙통제 장치가 결여 되어 있기 때문에 중앙통제 장치 개입을 최소로 하면서 P2P를 운영하기 위한 효율적인 기법 및 체계가 요구되고 있다. 본 논문에서는 기존의 접근방법을 검토하여 FFNN(feed forward neural network)을 이용한 디바이스 탐색 기법을 제시한다. 제시한 FFNN은 BP(back propagation) 알고리즘을 통해 훈련하고 디바이스를 탐색한다. 제시한 시스템의 성능을 보이기 위해 일정한 계산량을 가지는 작업을 에이전트를 활용, 탐색된 디바이스간에 분배하여 처리한다. 본 논문에서는 제한된 자원을 가지는 디바이스 간에 P2P를 사용하는 기법에 대해 제시하였다.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.9
• /
• pp.1700-1707
• /
• 2011

Interior permanent magnet synchronous motor(IPMSM) adjustable speed drives offer significant advantages over induction motor drives in a wide variety of industrial applications such as high power density, high efficiency, improved dynamic performance and reliability. Since the fuzzy neural network(FNN) is recognized general approximate method to control non-linearities and uncertainties, the development of FNN control systems have also grown rapidly. The FNN controller is compounded of fuzzy and neural network. It has an advantage that is the robustness of fuzzy control and the ability to adapt of neural network. However, the FNN has static problem due to their feed-forward network structure. This paper proposes high performance speed control of IPMSM drive using the recurrent FNN(RFNN) which improved conventional FNN controller. The RFNN has excellent dynamic response characteristics because of it has internally feed-back structure. Also, this paper proposes speed estimation of IPMSM drive using ANN. The proposed method is analyzed and compared to conventional FNN controller in various operating condition such as parameter variation, steady and transient states etc.

• Journal of Korean Society on Water Environment
• /
• v.21 no.1
• /
• pp.21-28
• /
• 2005

The training and prediction modeling using an artificial neural network was implemented to predict the turbidity of treated water as well as to estimate the optimized feed concentration of polyaluminium chloride (PACl) in a water treatment plant. The parameters used in the input layers were pH, temperature, turbidity and alkalinity, while those in output layers were PACl and turbidity of treated water. Levenberg-Marquadt method of feedforward back-propagation perceptron in the neural network toolbox of MATLAB program was used in this study. Correlation coefficients of the training data with the measured data were 0.9997 for PACl and 0.6850 for turbidity and those of the testing data with measured data were 0.9140 for PACl and 0.3828 for turbidity, when four parameters at input layer, 12-12 nodes each at both the first and the second hidden layers, and two parameters(PACl and turbidity) at output layer were used. Although the predictability of PACl was improved, compared to that of the previous studies to use the only coagulant dose as output layer, turbidity in treated water could not be predicted well. Acquisition of more data through several years obtained with the advanced on-line measuring system could make the artificial neural network useful and practical in actual water treatment plants.

• Korean Journal of Community Nutrition
• /
• v.1 no.3
• /
• pp.433-440
• /
• 1996

This study was carried out using a questionnaire in order to investigate factors affecting the status and the method of infant feeding in Incheon area. The subjects were 126 monthers of infants ranging from 3 months to 18 months. The results are : the percentages of those feeding colostrum and breast milk has decreased in comparison with the previous results. But the percentage of breast-feeding gets higher in a case where colostrum or breast milk is first fed after delivery. Also frequency and birth order of child plays a part : the more frequently monthers try to feed breast milk in the earliest period of feeding, the more they keep breast-feeding, and the latter-born(the second-born or the third born) children get more chances to be breast-fed than the first-born. Another noticeable factor is what babies feed on during hospitalization. if they get fed more breast milk than infant formula while they are in hospital, they tend to keep feeding on breast milk after they come back home. Breast-feeding group have good knowledge and information abut the importance of breast-feeding. The data shows that they believe the superiority of breast milk.

• Journal of Korean Society of Water and Wastewater
• /
• v.17 no.4
• /
• pp.519-527
• /
• 2003

Cake layer in crossflow microfiltration(CFMF) can be reduced by coagulation, enhancing membrane flux. This is because enlarging particle size by coagulation increases shear-induced diffusivity and the back-transport of rejected particles. However it is known that the enlarged particles are disaggregated by the shear force of the pump while passing through it. This study is to look at the disaggregation in relation with cake layer reducation. Kaolin and polysulfon hollow fiber microfilter are used for experiment. The reduction of cake resistance by coagulation is observed in a range of 17% to 53% at the various coagulation conditions. The particle size analysis results of the experiments show that aggregated particles in feed are completely disaggregated by pump but re-aggregation of particles occurs in membrane. This suggestes that the re-aggregation of particles is critical to cake reduction and flux enhancement, since the aggregated particles are completely broken. The mechanisms for re-aggregation in membrane are the same with those for coagulation in feed tank. Charge neutralization is better for CCFMF than sweep flocculation although it has two drawbacks in operation.

• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.3
• /
• pp.321-328
• /
• 1998

When we control the level of the steam generator in the nuclear power plants, a swell and shrink arises from many disturbances such as feed water rate, feed water temperature, main steam flow rate, and coolant temperature. If we use the conventional type of PI controller in this system, we will not have stability during controlling at lower power, the removal function of disturbances, and a load follow-up control effectively. In this paper, we study the application of a 2-Degree of Freedom(2-DOF) PID controller to the level control of the steam. generator of nuclear power plants through the simulation and the experimental steam generator. We use the parameters $\alpha$, $\beta$, $\gamma$ of the 2-DOF PID controller for the removal of disturbances and the parameters Kp,Ti,Td of the conventional type of PID controller for controlling setpoint. The back-propagation learning algorithm of neural network is used for tuning the 2-DOF PID controller. We can find satisfactory results of the removal of the disturbances and the tracking function in the change of setpoint through the simulation and experimental steam generator.

• 한국연소학회:학술대회논문집
• /
• 2015.12a
• /
• pp.131-134
• /
• 2015

국내 천연가스 열량제도를 현행 표준 열량제 $10,400kcal/Nm^3(43.54MJ/Nm^3)$에서 중간 조정기간을 두고 2012.07.01부터는 최저 $10,100kcal/Nm^3(42.28MJ/Nm^3)$을 유지하고 2015년 이후 $9,800(41.1MJ/Nm^3){\sim}10,600kcal/Nm^3(44.4MJ/Nm^3)$ 열량범위제도로 변경 추진되고 있다. 산업현장에서 열량변동을 측정하여 공연비 제어기술을 개발하고자 60ton Al 용해로에 열량대응기술 개발을 위한 내용으로 열량측정시스템설치 및 열량 값과 연계하여 공연비 제어기술개발연구 내용으로 결과는 다음과 같다. 단순히 표준열량으로 에 맞춰 프로그램된 제어로직에 열량변동에서 검출된 신호를 이용하여 연료보정 값을 추가한 로직을 재구성할 필요가 있다. 이 혀장의 경우는 용탕의 온도가 목표온도 근처까지 올리기가 어려워진 상황으로 주로 공급열량 저열량화에 따른 과잉공기영향으로 온도상승이 어려워 보이며 적절한 공연비로 최적화 되면 이러한 문제가 개선되리라 생각된다.