• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.215-216
• /
• 2007

This paper presents a real-time scheduler for nuclear safety system. According to constraints and requirements of nuclear safety system, scheduler design analysis is done and algorithms are developed for implementation. Using DSP based hardware, a real-time scheduler is realized. Consequently, this paper shows the performance of periodical software through the monitoring program.

• International conference on construction engineering and project management
• /
• 2009.05a
• /
• pp.27-29
• /
• 2009

Real-time control of on-site construction, based on high quality data, is essential to identify discrepancies between actual and planned performances. Additionally, real-time control enables timely corrective measures to be taken when needed to reduce the damages caused by the discrepancies. The focus of the presentation will be on our work, which uses automated data technologies to collect data needed for real time control.

• Spatial Information Research
• /
• v.17 no.3
• /
• pp.319-328
• /
• 2009

The growth of sensor and wireless communication technology enable us to monitor the status of things in real-time. There are many attempts to apply the real-time monitoring in various fields. Especially, the real-time monitoring in facility management fields has attracted much attentions recently. But it is difficult to apply recent technologies including sensor, wireless communication and Geospatial information technologies, etc, to the real situation directly. One of the reasons is that facility managers who are familiar with existing processes may resist the change induced by new technologies. So it is necessary to close the gap between new technologies and existing processes by improving the existing processes prior to putting them in the field. In this paper, we suggest the management process improvement focused on the district heating facilities.

• The KIPS Transactions:PartA
• /
• v.10A no.6
• /
• pp.715-722
• /
• 2003

Monitoring is used to see if a real-time system provides a service on time. The target of monitoring is not only an interior system but also a remote system which is located in the remote network. Monitoring needs data store to monitor data from each system. But a data store should be designed on the considerations of time constraints and data accessibility. In this paper, we present an architecture of XML-based data store and network delivery. XML-based data store is based on XML which is a standardized data format. So any platform which supports TCP/IP and HTTP can access data in the data store without any conversion. The XML-based delivery architecture is designed to reduce the time of data access and delivery. In addition, some experiments were tested to provide the timing guideline to be kept by a real-time system which uses the architecture presented in this paper. The architecture of XML-based data store and delivery designed in this paper can be used in the domains of remote real-time monitoring and control.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.19 no.3
• /
• pp.151-158
• /
• 2024

Although on-device artificial intelligence (AI) has gained attention to diagnosing machine faults in real time, most previous studies did not consider the model retraining and redeployment processes that must be performed in real-world industrial environments. Our study addresses this challenge by proposing an on-device AI-based real-time machine fault diagnosis system that utilizes continual learning. Our proposed system includes a lightweight convolutional neural network (CNN) model, a continual learning algorithm, and a real-time monitoring service. First, we developed a lightweight 1D CNN model to reduce the cost of model deployment and enable real-time inference on the target edge device with limited computing resources. We then compared the performance of five continual learning algorithms with three public bearing fault datasets and selected the most effective algorithm for our system. Finally, we implemented a real-time monitoring service using an open-source data visualization framework. In the performance comparison results between continual learning algorithms, we found that the replay-based algorithms outperformed the regularization-based algorithms, and the experience replay (ER) algorithm had the best diagnostic accuracy. We further tuned the number and length of data samples used for a memory buffer of the ER algorithm to maximize its performance. We confirmed that the performance of the ER algorithm becomes higher when a longer data length is used. Consequently, the proposed system showed an accuracy of 98.7%, while only 16.5% of the previous data was stored in memory buffer. Our lightweight CNN model was also able to diagnose a fault type of one data sample within 3.76 ms on the Raspberry Pi 4B device.

• Journal of Biomedical Engineering Research
• /
• v.16 no.2
• /
• pp.175-182
• /
• 1995

Despite the simplicity of processing, a conventional autocorrelation function (ACF) method for the precise determination of fetal heart rate (FHR) has many problems. In case of weak or noise corrupted Doppler ultrasound signal, the ACF method is very sensitive to the threshold level and data window length. It is very troublesome to extract FHR when there is a data loss. To overcome these problems, the high resolution pitch detection algorithm was adopted to estimate the FHR. This method is more accurate, robust and reliable than the ACF method. With a lot of calculation, however, it is impossible to process real time FHR estimation. This paper is presented a new FHR estimation algorithm for real time processing and studied the real time FHR monitoring system by high resolution pitch detection algorithm.

• Journal of Welding and Joining
• /
• v.16 no.5
• /
• pp.76-85
• /
• 1998

The resistance spot welding process has been used for joining the sheet metal in automotive engineering. In the resistance spot welding, the weld quality is very important, because the quality of weld is one of the most important factors to the automobile quality. The size of he molten nugget has been utilized to estimate the weld quality. However, it is not easy to find the weld defects. For weldability estimation, we have to use the nondestructive method such as X-ray or ultrasonic inspection. But these kinds of approaches are not suitable for detecting the defects in real time. The purpose of this study is to develop the real time monitoring of the weld quality in the resistance spot welding. Obtained data were used to estimate weldability using fuzzy algorithm. It is sound that this monitoring and estimation system can be useful to improve the weld quality in the resistance spot welding process and it is possible to estimate the weldability in real time.

• Food Science and Biotechnology
• /
• v.15 no.4
• /
• pp.595-598
• /
• 2006

Semi-quantitative monitoring of Lactobacillus sake and Lactobacillus plantarum, major and minor microorganisms in kimchi, respectively, and Lactobacillus paraplantarum, recently shown to be present in kimchi, was carried out by real-time polymerase chain reaction (PCR). Changes in the 3 species during kimchi fermentation were monitored by the threshold cycle ($C_T$) of real-time PCR. As fermentation proceeded at $15^{\circ}C$, the number of L. sake increased dramatically compared to those of L. plantarum and L. paraplantarum. During fermentation at $4^{\circ}C$, the growth rates of the 3 species decreased, but the proportions of L. plantarum and L. paraplantarum in the microbial ecosystem were almost constant. Considering the $C_T$ values of the first samples and the change in the $C_T$ value, the number of L. sake is no doubt greater than those of L. plantarum and L. paraplantarum in the kimchi ecosystem. L. sake seems to be one of the major microorganisms involved in kimchi fermentation, but there is insufficient evidence to suggest that L. plantarum is the primary acidifying bacterium.

• Journal of Positioning, Navigation, and Timing
• /
• v.11 no.4
• /
• pp.269-277
• /
• 2022

In this paper, we overview the system development status of the national maritime precise point positioning-real-time kinematic (PPP-RTK) service in Korea, also known as the Precise POsitioning and INTegrity monitoring (POINT) system. The development of the POINT service began in 2020, and the open service is scheduled to start in 2025. The architecture of the POINT system is composed of three provider-side facilities-a reference station, monitoring station, and central control station-and one user-side receiver platform. Here, we propose the detailed functionality of each component considering unidirectional broadcasting of augmentation data. To meet the centimeter-level user positioning accuracy in maritime coverage, new reference stations were installed. Each reference station operates with a dual receiver and dual antenna to reduce the risk of malfunctioning, which can deteriorate the availability of the POINT service. The initial experimental results of a testbed from corrections generated from the testbed network, including newly installed reference stations, are presented. The results show that the horizontal and vertical accuracies satisfy 2.63 cm and 5.77 cm, respectively. For the purpose of (near) real-time broadcasting of POINT correction data, we designed a correction message format including satellite orbit, satellite clock, satellite signal bias, ionospheric delay, tropospheric delay, and coordinate transformation parameters. The (near) real-time experimental setup utilizing (near) real-time processing of testbed network data and the designed message format are proposed for future testing and verification of the system.

• Journal of KIBIM
• /
• v.11 no.2
• /
• pp.24-32
• /
• 2021

Recently, starting with smart construction research, interest in technology that automates construction site management using artificial intelligence technology is increasing. In order to automate construction site management, it is necessary to recognize objects such as construction equipment or workers, and automatically analyze the relationship between them. For example, if the relationship between workers and construction equipment at a construction site can be known, various use cases of site management such as work productivity, equipment operation status monitoring, and safety management can be implemented. This study derives a real-time object detection platform architecture that is required when performing construction site management using deep learning technology, which has recently been increasingly used. To this end, deep learning models that support real-time object detection are investigated and analyzed. Based on this, a deep learning model development process required for real-time construction site object detection is defined. Based on the defined process, a prototype that learns and detects construction site objects is developed, and then platform development considerations and architecture are derived from the results.