• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.499-499
• /
• 2023

최근 컴퓨터 성능 향상과 새로운 머신러닝 알고리즘 개발됨에 따라, 각 분야별 연구자들이 이를 활용한 연구를 다양하게 수행하고 있으며, 하수처리시설의 경우에는 막대한 양의 운영자료가 축척됨에 따라 머신러닝을 활용한 다양한 연구가 가속화 되고 있다. 기존 하수처리장의 물리학적 모델은 적용된 영향 인자에 여러 가지 가정이 고려되어 모델 정확도가 부정확해지는 경향이 있었으며, 이러한 문제점을 보완하기 위해 하수처리장의 수집된 운영자료 및 머신러닝 기반의 예측 모델을 활용하여 예측 모델 정확도를 향상하는 선행 연구들이 진행되고 있다. A 하수처리장의 부지 내에 설치된 센서를 통하여 운영자료가 중앙제어실 서버에 실시간으로 저장되는 자료를 활용하여 NN (Neural Network), SVM (Support Vector Machine), RF (Random Forest) 등과 같은 다양한 머신러닝 모델을 적용하였고, 하수처리장 운영자료를 적용할 경우 어느 모델이 가장 높은 성능이 나타나는지 인사이트를 도출하고자 하였다. 금회 연구는 A 하수처리장을 대상으로 여러 머신러닝 기반 예측 모델을 개발하고, 각 모델의 예측정확도를 서로 평가함으로써, 머신러닝 모델 최적화를 수행할 수 있었다. 이번 연구에서 도출된 결과를 활용하여 하수처리장 예측 모델 최적화를 진행할 경우, 향후 비교적 짧은 시간에 하수처리장 머신러닝 기반 예측 모델 개발이 가능하다는 점에 의의가 있다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.185-188
• /
• 2021

본 논문에서는 기술의 빠른 발전에도 불구하고 줄어들지 않는 화재 사고, 그 중에서도 많은 인명피해를 내는 실내 화재 사고에 대하여 기존 실내 화재 알림 시스템의 한계점인 알림의 양치기 소년화로 인한 안전 불감증 증가와 알림의 사각지대 문제를 해결하고자 새로운 대안 시스템을 설계 및 구현하고, 실험 검증을 진행하였다. 위 두 가지 문제점을 해결하기 위해, 본 연구에서는 스마트폰이 매우 대중적으로 보급되어 있다는 점을 기반으로 IoT, 엣지 컴퓨팅, 비콘 등을 응용한다. 비콘 신호를 broadcasting 하는 엣지 노드의 신호 범위 내에 진입하면 사용자 정보를 수집하여 대상 건물에 출입한 대상을 특정한다. 말단 센서 노드와 엣지 노드 간의 무선 RF 통신으로 화재를 모니터링하며 화재가 발생했을 시 특정된 대상들에게만 스마트폰 어플의 푸시 알림으로 화재 발생 상황을 전송하는 시스템을 설계 및 구현하였다. 시스템 성능 평가를 위해 동국대학교 건물 내에서 수평, 수직으로 이동하며 실험을 진행하였고, 그 결과를 통해 대안 시스템의 성능과 한계를 분석하여 이를 실내 공간에 적용하기 위한 적합성을 평가하였다.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.4
• /
• pp.35-42
• /
• 2011

Link quality assessment is a crucial part of sensor network formation to stably operate large-scale wireless sensor networks (WSNs). A stability of path consisting of several nodes strongly depends on all link quality between pair of consecutive nodes. Thus it is very important to assess the link quality on the stage of building a routing path. In this paper, we present a link quality assessment method, Hybrid Link Quality Metric (HQLM), which uses both of LQI and RSSI from RF chip of sensor nodes to minimize set-up time and energy consumption for network formation. The HQLM not only reduces the time and energy consumption, but also provides complementary cooperation of LQI and RSSI. In order to evaluate the validity and efficiency of the proposed method, we measure PDR (Packet Delivery Rate) by exchanging multiple messages and then, compare PDR to the result of HQLM for evaluation. From the research being carried out, we can conclude that the HQLM performs better than either LQI- or RSSI-based metric in terms of recall, precision, and matching on link quality.

• Korean Journal of Applied Biomechanics
• /
• v.18 no.2
• /
• pp.59-64
• /
• 2008

D. W. KANG, J. S. CHOI, and G. R. TACK, A Study on Real-Time Sports Activity Classification & Monitoring Using a Tri-axial Accelerometer. Korean Jouranl of Sport Biomechanics, Vol. 18, No. 2, pp. 59-64, 2008. This study was conducted to study the real-time sports activity classification and monitoring using single waist mounted tri-axial accelerometer. This monitoring system detects events of sports activities such as walking, running, cycling, transitions between movements, resting and emergency event of falls. Accelerometer module was developed small and easily attachable on waist using wireless communication system which does not constrain sports activities. The sensor signal was transferred to PC and each movement pattern was classified using the developed algorithm in real-time environment. To evaluate proposed algorithm, experiment was performed with several sports activities such as walking, running, cycling movement for 100sec each and falls, transition movements(sit to stand, lie to stand, stand to sit, lie to sit, stand to lie and sit to lie) for 20 times each with 5 healthy subjects. The results showed that successful detection rate of the system for all activities was 95.4%. In this study, through sports activity monitoring. it was possible to classify accurate sports activities and to notify emergency event such as falls. For further study, the accurate energy consumption algorithm for each sports activity is under development.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.1
• /
• pp.1-9
• /
• 2009

In this paper, a 2.4 GHz doppler radar system consisting of a doppler radar sensor and a baseband module were designed to detect heart beat and respiration signal without direct skin contact. The doppler radar system emits RF signal of 2.4 GHz toward human chest, and then detects phase modulation of the reflected signal so as to investigate cardiopulmonary activities. The heartbeat and respiration signals acquired from I/Q channels of the doppler radar system are applied to the pre-processing circuit, the amplification circuit, and the offset circuit of the baseband module. The designed system was tested on mouse, rabbit and mankind, which have different range of heart rates and respiration signals, to evaluate detection accuracy of the system. ECG acquisition system and respiration transducer were used to generate the reference signal. In our experiments, a performance of detection were found to be high in the case that the subject stays still. In this paper, we confirmed that non-contact heart beat and respiration detection using the doppler radar has the possibility and limitation according to distance, cardiopulmonary activities, range of heart rates and respiration.

• Journal of Bio-Environment Control
• /
• v.31 no.3
• /
• pp.221-229
• /
• 2022

This study was aimed to investigate spatial and vertical characteristics of greenhouse environments according to the location of the environmental sensors, and to investigate the correlations between temperature, light intensity, and carbon dioxide (CO₂) concentration according to the type of greenhouse. Temperature, relative humidity (RH), CO₂, and light sensors were installed in the four-different vertical positions of the whole canopy as well as ground and roof space at the five spatial locations of the Venlo greenhouse. Also, correlations between temperature, light intensity, and CO₂ concentration in Venlo and semi-closed greenhouses were analyzed using the Curve Expert Professional program. The deviations among the spatial locations were larger in the CO₂ concentration than other environmental factors in the Venlo greenhouse. The average CO₂ concentration ranged from 465 to 761 µmol·mol^-1 with the highest value (646 µmol·mol^-1) at the Middle End (4ME) close to the main pipe (50Ø) of the liquefied CO₂ gas supply and lowest (436 µmol·mol^-1) at the Left Middle (5LM). The deviation among the vertical positions was greater in temperature and relative humidity than other environments. The time zone with the largest deviation in average temperature was 2 p.m. with the highest temperature (26.51℃) at the Upper Air (UA) and the lowest temperature (25.62℃) at the Lower Canopy (LC). The time zone with the largest deviation in average RH was 1 p.m. with the highest RH (76.90%) at the LC and the lowest RH (71.74%) at the UA. The highest average CO₂ concentration at each hour was Roof Air (RF) and Ground (GD). The coefficient of correlations between temperature, light intensity, and CO₂ concentration were 0.07 for semi-closed greenhouse and 0.66 for Venlo greenhouse. All the results indicate that while the CO₂ concentration in the greenhouse needs to be analyzed in the spatial locations, temperature and humidity needs to be analyzed in the vertical positions of canopy. The target CO₂ fertilization concentration for the semi-closed greenhouse with low ventilation rate should be different from that of general greenhouses.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.5
• /
• pp.219-224
• /
• 2015

In this paper, the designed fish robots DOMI ver1.0 is researched and development for aquarium underwater robot. The presented fish robot consists of the head, 1'st stage body, 2nd stage body and tail, which is connected two point driving joints. The model of the robot fish is analysis to maximize the momentum of the robot fish and the body of the robot is designed through the analysis of the biological fish swimming. Also, Lighthill was applied to the kinematics analysis of robot fish swimming algorithms, we are applied to the approximate method of the streamer model that utilizes techniques mimic the biological fish. The swimming robot has two operating mode such as manual and autonomous operation modes. In manual mode the fish robot is operated to using the RF transceiver, and in autonomous mode the robot is controlled by microprocessor board that is consist PSD sensor for the object recognition and avoidance. In order to the submerged and emerged, the robot has the bladder device in a head portion. The robot gravity center weight is transferred to a one-axis sliding and it is possible to the submerged and emerged of DOMI robot by the breath unit. It was verified by the performance test of this design robot DOMI ver1.0. It was confirmed that excellent performance, such as driving force, durability and water resistance through the underwater field test.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.8
• /
• pp.521-529
• /
• 2020

Forest fires are one of the most important environmental risks that have adverse effects on many aspects of life, such as the economy, environment, and health. The early detection, quick prediction, and rapid response of forest fires can play an essential role in saving property and life from forest fire risks. For the rapid discovery of forest fires, there is a method using meteorological data obtained from local sensors installed in each area by the Meteorological Agency. Meteorological conditions (e.g., temperature, wind) influence forest fires. This study evaluated a Data Mining (DM) approach to predict the burned area of forest fires. Five DM models, e.g., Stochastic Gradient Descent (SGD), Support Vector Machines (SVM), Decision Tree (DT), Random Forests (RF), and Deep Neural Network (DNN), and four feature selection setups (using spatial, temporal, and weather attributes), were tested on recent real-world data collected from Gyeonggi-do area over the last five years. As a result of the experiment, a DNN model using only meteorological data showed the best performance. The proposed model was more effective in predicting the burned area of small forest fires, which are more frequent. This knowledge derived from the proposed prediction model is particularly useful for improving firefighting resource management.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.15 no.2
• /
• pp.263-267
• /
• 2015

current mine detector can't division the section if it is conducted and it needs too much labor force and time. in addition to, if the user don't move the head of sensor in regular speed or move it too fast, it is hard to detect a mine exactly. according to this, to improve the problem using one direction ultrasonic wave sensing signal, that is made up of human body antenna part, main micro processor unit part, smart glasses part, body equipped LCD monitor part, wireless data transmit part, belt type power supply part, black box type camera, Security Communication headset. the user can equip this at head, body, arm, waist and leg in removable type. so it is able to detect the powder in a 360-degree on(under) the ground whether it is metal or nonmetal and it can express the 2D or 3D film about distance, form and material of the mine. so the battle combats can avoid the mine and move fast. also, through the portable battery and twin self power supply system of the power supply part, combat troops can fight without extra recharge and we can monitoring the battle situation of distant place at the command center server on real-time. and then, it makes able to sharing the information of battle among battle combats one on one. as a result, the purpose of this study is researching a smart wearable mine detector which can establish a smart battle system as if the commander is in the site of the battle.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.4
• /
• pp.55-62
• /
• 2004

In this paper, we have developed a portable acceleration measurement system to measure acceleration signals during walking and a gait analysis algorithm which can evaluate gait regularity and symmetry and estimate gait parameters automatically. Portable acceleration measurement system consists of a biaxial accelerometer, amplifiers, lowpass filter with cut-off frequency of 16Hz, one-chip microcontroller, EEPROM and RF(TX/RX) module. The algerian includes FFT analysis, filter processing and detection of main peaks. In order to develop the algorithm, eight hemiplegic patients for training set and the other eight hemiplegic patients for test set are participated in the experiment. Acceleration signals during 10m walking were measured at 60 samples/sec from a biaxial accelerometer mounted between L3 and L4 intervertebral area. The algorithm, detected foot contacts and classified right/left steps, and then calculated gait parameters based on these informations. Compared with video data and analysis by manual, algorithm showed good performance in detection of foot contacts and classification of right/left steps in test set perfectly. In the future, with improving the reliability and ability of the algerian so that calculate more gait Parameters accurately, this system and algerian could be used to evaluate improvement of walking ability in hemiplegic patients in clinical practice.