• ETRI Journal
• /
• v.42 no.4
• /
• pp.468-479
• /
• 2020

The proposed AI processor architecture has high throughput for accelerating the neural network and reduces the external memory bandwidth required for processing the neural network. For achieving high throughput, the proposed super thread core (STC) includes 128 × 128 nano cores operating at the clock frequency of 1.2 GHz. The function-safe architecture is proposed for a fault-tolerance system such as an electronics system for autonomous cars. The general-purpose processor (GPP) core is integrated with STC for controlling the STC and processing the AI algorithm. It has a self-recovering cache and dynamic lockstep function. The function-safe design has proved the fault performance has ASIL D of ISO26262 standard fault tolerance levels. Therefore, the entire AI processor is fabricated via the 28-nm CMOS process as a prototype chip. Its peak computing performance is 40 TFLOPS at 1.2 GHz with the supply voltage of 1.1 V. The measured energy efficiency is 1.3 TOPS/W. A GPP for control with a function-safe design can have ISO26262 ASIL-D with the single-point fault-tolerance rate of 99.64%.

• Journal of Digital Contents Society
• /
• v.14 no.2
• /
• pp.143-149
• /
• 2013

Many researchers and developers take a greater interest on the LOD AI techniques as users demand more elaborate and sophisticated game AI in recent years. However, contrary to the traditional geometry LOD, existing LOD AI techniques can be used only to a limited extent. Therefore, in this paper, I propose an LOD AI technique, which uses HSM(Hierarchical State Machine) and the Lua script language as the method to control game objects. Using the proposed approach, we can easily produce multilevel AI models for LOD and design various objects without hard-coding state machines. Moreover, in order to show the effectiveness of the presented technique, this paper exemplifies the results of the efficiency test through the prototype engine.

• Journal of Practical Engineering Education
• /
• v.13 no.2
• /
• pp.301-304
• /
• 2021

In this study, we developed a module that can perform two functions at the same time through interworking between a personal recognition module and a heat detection module in the era of COVID-19. This can simultaneously solve the problem of compatibility of the personal recognition module that occurs in the existing system and the problem of secondary infection that can occur during congestion due to the separate implementation of heat detection. Therefore, in this study, NFC and Bluetooth motherboards were developed, and an array-type non-contact temperature sensor was applied to detect heat. The developed system is expected to be able to realize both access control of floating population and effective quarantine at the same time in public institutions or private companies that require AI automated access control. In addition, it is judged that it is possible to link the embedded programming and web programming implementation method using the module of the development system to the educational program.

• Journal of Information Technology Services
• /
• v.20 no.1
• /
• pp.29-40
• /
• 2021

North Korea's science and technology policies are being pursued under strong leadership and control by the central government. In particular, a large part of the research and development of science and technology related to the Fourth Industrial Revolution in North Korea is controlled and absorbed by the defense organizations under the national defense-oriented policy framework, among which North Korea is making national efforts to develop advanced technologies in artificial intelligence and actively utilize them in the military affairs. The future weapon system based on AI will have superior performance and destructive power that is different from modern weapons systems, which is likely to change the paradigm of the future battlefield, so a thorough analysis and prediction of the level of AI militarization technology, the direction of development, and AI-based weapons system in North Korea is needed. In addition, research and development of South Korea's corresponding weapon systems and military science and technology are strongly required as soon as possible. Therefore, in this paper, we will analyze the level of AI technology, the direction of AI militarization, and the AI-based weapons system in North Korea, and discuss the AI military technology and corresponding weapon systems that South Korea military must research and develop to counter the North Korea's. The next study will discuss the analysis of AI militarization technologies not only in North Korea but also in neighboring countries in Northeast Asia such as China and Russia, as well as AI weapon systems by battlefield function, detailed core technologies, and research and development measures.

• Journal of Korea Water Resources Association
• /
• v.57 no.3
• /
• pp.151-164
• /
• 2024

In water treatment plants supplying potable water, the management of chlorine concentration in water treatment processes involving pre-chlorination or intermediate chlorination requires process control. To address this, research has been conducted on water quality prediction techniques utilizing AI technology. This study developed an AI-based predictive model for automating the process control of chlorine disinfection, targeting the prediction of residual chlorine concentration downstream of sedimentation basins in water treatment processes. The AI-based model, which learns from past water quality observation data to predict future water quality, offers a simpler and more efficient approach compared to complex physicochemical and biological water quality models. The model was tested by predicting the residual chlorine concentration downstream of the sedimentation basins at Plant, using multiple regression models and AI-based models like Random Forest and LSTM, and the results were compared. For optimal prediction of residual chlorine concentration, the input-output structure of the AI model included the residual chlorine concentration upstream of the sedimentation basin, turbidity, pH, water temperature, electrical conductivity, inflow of raw water, alkalinity, NH₃, etc. as independent variables, and the desired residual chlorine concentration of the effluent from the sedimentation basin as the dependent variable. The independent variables were selected from observable data at the water treatment plant, which are influential on the residual chlorine concentration downstream of the sedimentation basin. The analysis showed that, for Plant, the model based on Random Forest had the lowest error compared to multiple regression models, neural network models, model trees, and other Random Forest models. The optimal predicted residual chlorine concentration downstream of the sedimentation basin presented in this study is expected to enable real-time control of chlorine dosing in previous treatment stages, thereby enhancing water treatment efficiency and reducing chemical costs.

• ETRI Journal
• /
• v.42 no.4
• /
• pp.505-517
• /
• 2020

The increasing size and complexity of deep neural networks (DNNs) necessitate the development of efficient high-performance accelerators. An efficient memory structure and operating scheme provide an intuitive solution for high-performance accelerators along with dataflow control. Furthermore, the processing of various neural networks (NNs) requires a flexible memory architecture, programmable control scheme, and automated optimizations. We first propose an efficient architecture with flexibility while operating at a high frequency despite the large memory and PE-array sizes. We then improve the efficiency and usability of our architecture by automating the optimization algorithm. The experimental results show that the architecture increases the data reuse; a diagonal write path improves the performance by 1.44× on average across a wide range of NNs. The automated optimizations significantly enhance the performance from 3.8× to 14.79× and further provide usability. Therefore, automating the optimization as well as designing an efficient architecture is critical to realizing high-performance DNN accelerators.

• Journal of Veterinary Science
• /
• v.23 no.3
• /
• pp.36.1-36.14
• /
• 2022

Background: Since 2003, the H5 highly pathogenic avian influenza (HPAI) subtype has caused massive economic losses in the poultry industry in South Korea. The role of inland water bodies in avian influenza (AI) outbreaks has not been investigated. Identifying water bodies that facilitate risk pathways leading to the incursion of the HPAI virus (HPAIV) into poultry farms is essential for implementing specific precautionary measures to prevent viral transmission. Objectives: This matched case-control study (1:4) examined whether inland waters were associated with a higher risk of AI outbreaks in the neighboring poultry farms. Methods: Rivers, irrigation canals, lakes, and ponds were considered inland water bodies. The cases and controls were chosen based on the matching criteria. The nearest possible farms located within a radius of 3 km of the case farms were chosen as the control farms. The poultry farms were selected randomly, and two HPAI epidemics (H5N8 [2014-2016] and H5N6 [2016-2017]) were studied. Conditional logistic regression analysis was applied. Results: Statistical analysis revealed that inland waters near poultry farms were significant risk factors for AI outbreaks. The study speculated that freely wandering wild waterfowl and small animals contaminate areas surrounding poultry farms. Conclusions: Pet birds and animals raised alongside poultry birds on farm premises may wander easily to nearby waters, potentially increasing the risk of AI infection in poultry farms. Mechanical transmission of the AI virus occurs when poultry farm workers or visitors come into contact with infected water bodies or their surroundings. To prevent AI outbreaks in the future, poultry farms should adopt strict precautions to avoid contact with nearby water bodies and their surroundings.

• Journal of Digital Policy
• /
• v.1 no.2
• /
• pp.1-8
• /
• 2022

In case of a dangerous situation, the roof, which serves as an emergency exit, must be open in case of fire according to the Fire Act. However, when the roof door is opened, it has become a place of various incidents and accidents such as illegal entry, crime, and suicide. As a result, it is a reality to close the roof door in terms of facility management to prevent crime, various incidents, and accidents. Accordingly, the government is pushing to legislate regulations on housing construction standards, etc. that mandate the installation of electronic automatic opening and closing devices on rooftop doors. Therefore, in this paper, an intelligent emergency door opening/closing device system is proposed. To this end, an intelligent emergency door opening and closing system was developed by linking wired and wireless access counting and AI image analysis. Finally, it is possible to build a wireless communication-based integrated management platform that provides remote control and history management in a centralized method of device status real-time monitoring and event alarm.

• Journal of Internet Computing and Services
• /
• v.23 no.6
• /
• pp.97-105
• /
• 2022

This paper explains trends in security technologies for ICS. Since ICS is usually applied to large-scale national main infrastructures and industry fields, minor errors caused by cyberattack could generate enormous economic cost. ICS has different characteristic with commonly used IT systems, so considering security threats of ICS separately with IT is needed for developing modern security technology. This paper introduce framework for ICS that analyzes recent cyberattack tactics & techniques and find out trends in Intrusion Detection System (IDS) which is representative technology for ICS security, and analyzes AI technologies used for IDS. Specifically, this paper explains data collection and analysis for applying AI techniques, AI models, techniques for evaluating AI Model.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.1-8
• /
• 2024

On-device AI technology, which can operate AI models at the edge devices to support real-time processing and privacy enhancement, is attracting attention. As intelligent IoT is applied to various industries, services utilizing the on-device AI technology are increasing significantly. However, general deep learning models require a lot of computational resources for inference and learning. Therefore, various lightweighting methods such as quantization and pruning have been suggested to operate deep learning models in embedded edge devices. Among the lightweighting methods, we analyze how to lightweight and apply deep learning models to edge computing devices, focusing on pruning technology in this paper. In particular, we utilize dynamic and static pruning techniques to evaluate the inference speed, accuracy, and memory usage of a lightweight AI vision model. The content analyzed in this paper can be used for intelligent video control systems or video security systems in autonomous vehicles, where real-time processing are highly required. In addition, it is expected that the content can be used more effectively in various IoT services and industries.