• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.3
• /
• pp.419-432
• /
• 2019

Most of deep learning model training was proceeded by supervised learning, which is to train labeling data composed by inputs and corresponding outputs. Labeling data was directly generated manually, so labeling accuracy of data is relatively high. However, it requires heavy efforts in securing data because of cost and time. Additionally, the main goal of supervised learning is to improve detection performance for 'True Positive' data but not to reduce occurrence of 'False Positive' data. In this paper, the occurrence of unpredictable 'False Positive' appears by trained modes with labeling data and 'True Positive' data in monitoring of deep learning-based CCTV accident detection system, which is under operation at a tunnel monitoring center. Those types of 'False Positive' to 'fire' or 'person' objects were frequently taking place for lights of working vehicle, reflecting sunlight at tunnel entrance, long black feature which occurs to the part of lane or car, etc. To solve this problem, a deep learning model was developed by simultaneously training the 'False Positive' data generated in the field and the labeling data. As a result, in comparison with the model that was trained only by the existing labeling data, the re-inference performance with respect to the labeling data was improved. In addition, re-inference of the 'False Positive' data shows that the number of 'False Positive' for the persons were more reduced in case of training model including many 'False Positive' data. By training of the 'False Positive' data, the capability of field application of the deep learning model was improved automatically.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.2
• /
• pp.135-141
• /
• 2019

Internet of thing becomes more active, many sensor devices are increasing. Sensors can use network wired network or use mobile communication network. From the viewpoint of the transmission rate, the mobile communication network can be roughly divided into two types of high-speed communication and low-speed communication. In the case of hundreds of millions of sensors in the mobile communication network, resources are wasted to use high-speed communication. Communication is required to reduce the transmission rate and appropriately allocate resources without wasting such resources. As the Internet of Thing has been activated, Narrowband Internet of Thing(NB-IoT), which is one of the low-power technologies in recent mobile communications, is in the spotlight from various companies. Currently, it can be seen that only NB-IoT or other low power consumption communication has the potential to be able to connect to the Internet with rapidly increasing sensor devices. In this paper, we designed and implemented a heater controller using Huawei NB-IoT communication Module, a server that collects controller information, and an application that allows default settings for devices. The main function of this system is to collect temperature and heater status and give it to the server, control the heater from the server, and set parameters for the heater to operate automatically. The system can be applied to places where wired communication is not established, such as road information, smart agriculture, and small reservoirs as well as heaters.

• PNF and Movement
• /
• v.17 no.1
• /
• pp.81-92
• /
• 2019

Purpose: This study was an investigation of the effect of unstable supporting exercise in young adults with functional ankle instability. The study tested the use of a jumper and virtual reality (VR) training via a VR-head mounted display (HMD) system to provide functional improvement in proprioception, range of motion (ROM), ankle muscle strength, agility, and balance. Methods: The subjects comprised 61 young adults (in their twenties) with functional ankle instability to decide as less than 24 points using Cumberland ankle instability tool. The subjects were divided into three groups: VUS (VR-HMD and unstable supporting exercise, n = 20), VSS (VR-HMD and stable supporting exercise, n = 19), and NUS (non-VR-HMD and unstable supporting exercise, n = 22). The exercise program was conducted three times per week for three weeks. VR training via a VR-HMD system and a VR application on a smart mobile device were used with the VUS and VSS groups, and unstable supporting exercise was used in the VUS and NUS groups for 30 minutes. Proprioception, ROM, ankle muscle strength, agility, and balance were measured before and after training. Results: The VUS group showed significant differences in most results, including proprioception, ROM, ankle muscle strength, agility, and balance to compare before and after, and the VSS and NUS groups partially. Moreover, the VUS group had significant differences in most results when compared with the other groups. Conclusion: Unstable supporting exercise and VR training via a VR-HMD system improved functional ankle instability in terms of proprioception, ROM, ankle muscle strength, agility, and balance.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.3
• /
• pp.485-492
• /
• 2021

The wind turbines with a rated capacity of 50kW or less are generally considered as small class. Small wind turbines are an attractive alternative for off-grid power system and electric home appliances, both as stand-alone application and in combination with other energy technologies such as energy storage system, photovoltaic, small hydro or diesel engines. The research objective is to develop the 50kW scale wind turbine blades in ways that resemble as closely as possible with the construction and methods of utility scale turbine blade manufacturing. The mold process based on wooden form is employed to create a hollow, multi-piece, lightweight design using carbon fiber and fiberglass with an epoxy based resin. A hand layup prototyping method is developed using high density foam molds that allows short cycle time between design iterations of aerodynamic platforms. A production process of five blades is manufactured and key components of the blade are tested by IEC 61400-23 to verify the appropriateness of the design. Also, wind system with developed blades is tested by IEC 61400-12 to verify the performance characteristics. The results of blade and turbine system test showed the available design conditions for commercial operation.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.3
• /
• pp.268-278
• /
• 2021

Technologies that are leading the fourth industrial revolution, such as the Internet of Things (IoT), big data, artificial intelligence (AI), and cyber-physical systems (CPS) are developing and generalizing. The demand to improve productivity, economy, safety, etc., is spreading in various industrial fields by applying these technologies. Digital twins are attracting attention as an important technology trend to meet demands and is one of the top 10 tasks of the Korean version of the New Deal. In this study, papers, magazines, reports, and other literature were searched using Google. In order to investigate the contribution or role of geospatial information in the digital twin application, the definition of a digital twin, we investigated technology trends of domestic and foreign companies; the components of digital twins required in manufacturing, plants, and smart cities; and the core techniques for driving a digital twin. In addition, the contributing contents of geospatial information were summarized by searching for a sentence or word linked between geospatial-related keywords (i.e., Geospatial Information, Geospatial data, Location, Map, and Geodata and Digital Twin). As a result of the survey, Geospatial information is not only providing a role as a medium connecting objects, things, people, processes, data, and products, but also providing reliable decision-making support, linkage fusion, location information provision, and frameworks. It was found that it can contribute to maximizing the value of utilization of digital twins.

• Food Science and Industry
• /
• v.54 no.2
• /
• pp.93-101
• /
• 2021

Hazard Analysis Critical Control Point (HACCP) systems were developed to ensure a high level of food safety and reduced risk of foodborne illness. This paper focuses on significant issues associated with the implementation of HACCP; it provides an overview on recent literature. The structure of the paper follows six groupings of issues in the international literature of HACCP: (1) comparative studies and unification plan between HACCP and other food safety regulations; (2) verification of the HACCP system's effectiveness in improving food safety; (3) establishment of critical control point (CCP) for various foods HACCP model development; (4) expansion of HACCP application in the various fields and small businesses;(5) the impacts of HACCP on consumer's preferences and firms' financial performance in food industry; (6) HACCP and technological changes. The paper concludes with some suggestions for the future research in order to promote safe food supply chain for global customers.

• Journal of Digital Convergence
• /
• v.18 no.12
• /
• pp.29-36
• /
• 2020

The background of this study is that machine learning administrative services are recently attracting attention as a major policy tool for non-face-to-face administrative services in the post-corona era. This study investigated the types of work expected to be effective when introducing machine learning administrative services for Seoul Metropolitan Government officials who are piloting machine learning administrative services. The research method is a machine that can be introduced by organizational unit by distributing and collecting questionnaires for Seoul administrative organizations that have performed machine learning-based administrative services for one month in July 2020 targeting Seoul public officials using machine learning-based administrative services. By analyzing the learning administration service and application service, the business characteristics of each machine learning administration service type such as supervised learning work type, unsupervised learning work type, and reinforced learning work type were analyzed. As a result of the research analysis, it was found that there were significant differences in the characteristics of administrative tasks by supervised and unsupervised learning areas. In particular, it was found that the reinforcement learning domain contains the most appropriate business characteristics for machine learning administrative services. Implications were drawn. The results of this study can be provided as a reference material to practitioners who want to introduce machine learning administration services, and can be used as basic data for research to researchers who want to study machine learning administration services in the future.

• Journal of Sensor Science and Technology
• /
• v.30 no.3
• /
• pp.154-164
• /
• 2021

In this study, the IoT sensor technology required for improving the survival rate and high-density productivity of individual shrimp in smart shrimp farming (which involves the usage of recirculating aquaculture systems and biofloc technology) was analyzed. The principles and performances of domestic and overseas water quality monitoring IoT sensors were compared. Furthermore, the drawbacks of existing aquaculture monitoring technologies and the countermeasures for future aquaculture monitoring technologies were examined. In particular, for farming white-legged shrimp, an IoT sensor was employed to collect measurement indicators for managing the water quality environment in real-time, and the IoT sensor-based real-time monitoring technology was then analyzed for implementing the optimal farming environment. The results obtained from this study can potentially contribute to the realization of an autonomous farming platform that can improve the survival rate and productivity of shrimp, achieve feed reduction, improve the water quality environment, and save energy.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.13 no.6
• /
• pp.496-502
• /
• 2020

The electronic attendance management system is being introduced and operated on a pilot basis by some universities and educational institutions. However, most of the related systems have installed and operated the existing barcode and magnetic card systems. Classroom attendance is managed by introducing RF cards, but it causes problems such as recognition distance (less than 5cm) and the need for a check process in which students have to read the card each time with a reader for attendance. Also, it is not possible to respond in real time to the situation of midterm (early leave, absence from the second lecture time, etc.) because it is used in the lecture time of one subject with the record checked once. In order to solve these problems, the various mobile attendance systems proposed to solve these problems are also unable to fundamentally solve problems such as interim attendance and proxy attendance because they check attendance using only the application of a smartphone. In this paper, we use geofencing technology, which is a positioning-based technology that detects the entry and exit of people, objects, etc. in areas separated by virtual boundaries. The proposed system solves the problem of intermediate attendance and alternate attendance by setting the student to automatically record the access record when entering and leaving the classroom set as a geofence with a smartphone. In addition, it also provides a function to prevent unintentional mistakes that occur through the smartphone by limiting some of the functions of the smartphone such as silence, vibration, and Internet use when entering the classroom.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.133-142
• /
• 2021

With the recent advancements of deep learning, companies such as smart home, healthcare, and intelligent transportation systems are utilizing its functionality to provide high-quality services for vehicle detection, emergency situation detection, and controlling energy consumption. To provide reliable services in such sensitive systems, deep learning models are required to have high accuracy. In order to develop a deep learning model for analyzing previously mentioned services, developers should utilize the state of the art deep learning models that have already been verified for higher accuracy. The developers can verify the accuracy of the referenced model by validating the model on the dataset. For this validation, the developer needs structural information to document and apply deep learning models, including metadata such as learning dataset, network architecture, and development environments. In this paper, we propose a description language that represents the network architecture of the deep learning model along with its metadata that are necessary to develop a deep learning model. Through the proposed description language, developers can easily verify the accuracy of the referenced deep learning model. Our experiments demonstrate the application scenario of a deep learning description document that focuses on the license plate recognition for the detection of illegally parked vehicles.