• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.9
• /
• pp.1158-1165
• /
• 2021

This paper proposes a CNN accelerator which is optimized Pooling layer operation incorporated in Multiplication And Accumulation(MAC) to reduce the memory size. For optimizing memory and data path circuit, the quantized 8bit integer weights are used instead of 32bit floating-point weights for pre-training of MNIST data set. To reduce chip area, the proposed CNN model is reduced by a convolutional layer, a 4*4 Max Pooling, and two fully connected layers. And all the operations use specific MAC with approximation adders and multipliers. 94% of internal memory size reduction is achieved by simultaneously performing the convolution and the pooling operation in the proposed architecture. The proposed accelerator chip is designed by using TSMC65nmGP CMOS process. That has about half size of our previous paper, 0.8*0.9 = 0.72mm². The presented CNN accelerator chip achieves 94% accuracy and 77us inference time per an MNIST image.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.29A no.2
• /
• pp.100-106
• /
• 1992

In this pape, we designed and implemented a hardware accelerator for the generation of bit map font from Hangul outline font description for LBP (Laser Beam Printer) and screen applications Whole system was implemented as a double size PC/AT application board which consists of processing bolck and display block. The processing block has a master processor (MC68000)and two slave processors which are MC56001 and KAFOG chip responsible for the short vector generation. In the display block, TMS34061 was used for monitor display and GP425 was used for LBP print out. The resolution of the monitor is 640$\times$480 and that of LBP is 2385$\times$3390. The current system called KHGB90-B generates about 100 characters per second where each character consists of 32$\times$32 bits

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.1D
• /
• pp.81-88
• /
• 2012

This present research was carried out by dividing two parts; field surveying and data processing, in order to analyze changed patterns of a shoreline. Firstly, the shoreline information measured by the precise GPS positioning during long duration was collected. Secondly, the algorithm for detecting an auto boundary with regards to the changed shoreline with multi-image data was developed. Then, a comparative research was conducted. Haeundae beach which is one of the most famous ones in Korea was selected as a test site. RTK-GPS surveying had been performed overall eight times from September 2005 to September 2009. The filed test by aerial Lidar was conducted twice on December 2006 and March 2009 respectively. As a result estimated from both sensors, there is a slight difference. The average length of shoreline analyzed by RTK-GPS is approximately 1,364.6 m, while one from aerial Lidar is about 1,402.5 m. In this investigation, the specific algorithm for detecting the shoreline detection was developed by Visual C++ MFC (Microsoft Foundation Class). The analysis result estimated by aerial photo and satellite image was 1,391.0 m. The level of reliability was 98.1% for auto boundary detection when it compared with real surveying data.

• Journal of Astronomy and Space Sciences
• /
• v.18 no.2
• /
• pp.129-136
• /
• 2001

The accuracy of GPS applications is heavily dependent on the satellite ephemeris and earth orientation parameter. Specially applications like as the real time monitoring of troposphere and ionosphere require real time or predicted ephemeris arid earth orientation parameter with very high quality. IGS is producing IGS ultra rapid product called IGU for real time applications which includes the information of ephemeris and earth orientation. IGU is being made available twice everyday at 3:00 and 15:00 UTC arid covers 48 hours. The first 24 hours of it are based on actual GPS observations and the second 24 hours extrapolated. We will construct the processing strategy for yielding ultra rapid product and demonstrate the propriety through producing it using 48 hours data of 32 stations.

• Current Industrial and Technological Trends in Aerospace
• /
• v.8 no.1
• /
• pp.77-85
• /
• 2010

The current GPS constellation consists of 32 Block IIA/IIR/IIR-M satellites including 12 Block IIA satellites on service over 15 years. The satellites in poor space conditions may suffer from anomalies, especially influenced by aging atomic clocks which are of importance positioning and timing. Recently, the IGS Ultra-rapid predicted products have not shown acceptably high quality prediction performance because the Block IIA cesium clocks may be easily affected by various factors such as temperature and environment. The anomalies of aging clocks involve lower performance of positioning in the GPS applications. We, thus, describe satellite clock behaviors and anomalies induced by aging clocks and their detection technologies to avoid such anomalies.

• Proceedings of the Korean Society of Veterinary Pathology Conference
• /
• 2003.10a
• /
• pp.32-32
• /
• 2003

Intestinal infections are common in growing pigs and can be caused by multiple pathogens, environmental and management factors [1]. Among the most important viruses in swine enteritis are porcine epidemic diarrhea virus (PEDV), transmissible gastroenteritis virus (TGEV), porcine enteric calicivirus (PECV), porcine group A rotavirus (PRV gp A) and bacteria are Escherichia coli and Salmonella spp. and protozoa is Isospora suis [1]. The DNA chip system can serve as a powerful tool that can be utilized for simultaneous detection of specific pathogenic bacteria strains and viruses [2,3]. The combination of PCR and DNA chip technology will provide a novel method for the detection of porcine enteric pathogens thus revolutionize the diagnosis and management of the disease. The aim of this study is to develop DNA chip system for the rapid and reliable detection of five major porcine enteric pathogens based on oligonucleotide DNA chip hybridization. (omitted)

• Food Science of Animal Resources
• /
• v.32 no.4
• /
• pp.409-413
• /
• 2012

This study aimed to compare the probiotic characteristics of twelve strains of Lactobacillus acidophilus including cholesterol assimilation and conjugated linoleic acid (CLA) production. Cholesterol assimilation exhibited some variation among L. acidophilus strains, which could be classified into three groups based on their assimilation levels (p<0.05). The high cholesterol assimilation group exhibited a significantly higher tolerance to 0.3 and 0.5% bile acid than the low cholesterol assimilation group (p<0.05). Cholesterol assimilation showed positive correlation with 0.5% bile tolerance, and a negative correlation with acid tolerance (p<0.01). Glycocholate deconjugation activity showed no relationship with cholesterol assimilation, whereas taurocholate deconjugation activity was shown to have negative correlation with cholesterol assimilation (p<0.05). CLA production by L. acidophilus strains exhibited a wide variation, ranging from 2.69 to 5.04 mg/g fat. CLA production of L. acidophilus GP1B was the highest among the tested strains, but there was no evidence for differences in CLA production in strain specificity. Based on these results, the cholesterol assimilation of L. acidophilus strains may not be related to deconjugation activity, but may in-fact be attributed to their bile-tolerance.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.3C
• /
• pp.314-322
• /
• 2007

This paper proposes a new optimization algorithm named by GBNSGA(Goal-Pareto Based Non-dominated Sorting Genetic Algorithm) which determines the best configuration for CR(Cognitive Radio) communication systems. Conventionally, in order to select the proper radio configuration, genetic algorithm has been introduced so as to alleviate computational burden along the execution of the cognition cycle proposed by Mitola. This paper proposes a novel optimization algorithm designated as GBNSGA for cognitive engine which can be described as a hybrid algorithm combining well-known Pareto-based NSGA(Non-dominated Sorting Genetic Algorithm) as well as GP(Goal Programming). By conducting computer simulations, it will be verified that the proposed method not only satisfies the user's service requirements in the form of goals. It reveals the fast optimization capability and more various solutions rather than conventional NSGA or weighted-sum approach.

• Proceedings of the KIPE Conference
• /
• 2001.10a
• /
• pp.635-638
• /
• 2001

Photovoltaic (PV) pumping systems with maximum power point tracking (MPPT) technique aims at obtaining the highest possible power to the pump under various insolation and temperature, thus overcomes the mismatch between the photovoltaic panel and the pumping load. A simple method of tracking the maximum power points and forcing the system to operate close to these points is presented in this paper. The MC68HC908GP32 micro control unit (MCU) is employed to implement the proposed MPPT controller. Experimental results will also show the performances of the photovoltaic pumping system with the MPPT technique.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06a
• /
• pp.29-32
• /
• 2011

최근 GPS를 탑재한 스마트 기기의 보급으로 인해 사용자의 위치 정보를 이용한 위치기반서비스(LBS)가 많은 관심 속에 급속도로 성장하고 있다. 대부분의 위치기반서비스는 사용자의 위치를 바탕으로 주변의 유용한 정보를 제공하고 있으며 이러한 사용자 주변의 유용한 정보를 효율적으로 검색하기 위해서는 공간 데이터에 대한 색인 기법이 필요하다. 그러나 기존 공간 색인은 대부분 중앙 집중식 환경에서 연구가 진행되었고 분산 스토리지 환경에 대한 연구는 미흡하다. 따라서 본 연구는 서비스 성장과 더불어 서비스 영역이 확대됨에 따른 대용량 공간 데이터 환경에서 효율적인 공간 데이터 검색을 지원하는 그리드 기반 분산 색인 기법을 제안한다.