• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.7
• /
• pp.40-48
• /
• 2015

An optical network-on-chip(ONoC) architecture is emerging as a new paradigm for solving on-chip communication bottleneck. Recent studies on ONoC have been focusing on supporting the parallel transmission and avoiding path collisions using wavelength division multiplexing(WDM). However, since the maximum number of wavelengths, which a single waveguide can accommodate is limited by crosstalk and insertion loss. Therefore previous WDM studies based on incrementing the number of different wavelengths according to the number of nodes would be infeasible due to the implementation complexity. To solve such problems, we combined time division multiplexing(TDM) and wavelength-routed ONoC, along with an optimized channel allocation algorithm, which can minimize the number of extra wavelength channels and latency caused by combining TDM scheme.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.17 no.3
• /
• pp.467-472
• /
• 2022

The 3D ultrasonic anemometer transmits measured data by connecting PC and RS232C interface. Depending on the observation location, it is often necessary to extend the cable connecting the PC and the sensor. When installing on the test bed of the Air Meteorological Agency, the original AWM2919 cable was required to be extended because the distance between the PC container and the equipment installation site was more than 30 m. The cable was extended through a process such as extending the AWM2919 cable, converting the interface with the PC from RS232C to RS485, and testing the RS485 communication. After the equipment was installed with an extended cable, data were remotely collected and analyzed to confirm successful cable extension.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.6
• /
• pp.1-12
• /
• 2024

Deep learning with faster and more accurate results requires large amounts of storage space and large computations. Accordingly, many studies are using hardware accelerators for quick and accurate calculations. However, the performance bottleneck is due to data movement between the hardware accelerators and the CPU. In this paper, we propose a data prefetch strategy that can efficiently reduce such operational bottlenecks. The core idea of the data prefetch strategy is to predict the data needed for the next task and upload it to local memory while the hardware accelerator (Matrix Multiplication Unit, MMU) performs a task. This strategy can be enhanced by using a dual buffer to perform read and write operations simultaneously. This reduces latency and execution time of data transfer. Through simulations, we demonstrate a 24% improvement in the performance of hardware accelerators by maximizing parallel processing with dual buffers and bottlenecks between memories with data prefetch.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.45 no.4
• /
• pp.1-6
• /
• 2008

In this paper, millimeter-wave broadband MHEMT (Metamorphic High Electron Mobility Transistor) cascode amplifiers were designed and fabricated. The $0.1{\mu}m$ InGaAs/InAlAs/GaAs MHEMT was fabricated for cascode amplifiers. The DC characteristics of MHEMT are 670 mA/mm of drain current density, 588 mS/mm of maximum transconductance. The current gain cut-off frequency($f_T$) is 139 GHz and the maximum oscillation frequency($f_{max}$) is 266 GHz. To prevent oscillation of the designed cascode amplifiers, a parallel resistor and capacitor were connected to the drain of common gate device. By using the CPW (Coplanar Waveguide) transmission line, the cascode amplifier was designed and matched for the broadband characteristics. The designed amplifier was fabricated by the MHEMT MMIC process that was developed through this research. As the results of measurement, the amplifier was obtained 3 dB bandwidth of 50.37 GHz between 20.76 to 71.13 GHz. Also, this amplifier represents the S21 gain with the average 7.07 dB gain in bandwidth and the maximum gain of 10.3 dB at 30 GHz.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.7
• /
• pp.1924-1937
• /
• 1996

In this paper, we propose a ESP(Extended Scan Path) architecture for multi- board testing. The conventional architectures for board testing are single scan path and multi-scan path. In the single scan path architecture, the scan path for test data is just one chain. If the scan path is faulty due to short or open, the test data is not valid. In the multi-scan path architecture, there are additional signals in multi-board testing. So conventional architectures are not adopted to multi-board testing. In the case of the ESP architecture, even though scan paths either short or open, it doesn't affect remaining other scan paths. As a result of executing parallel BIST and IEEE 1149.1 boundary scan test by using, he proposed ESP architecture, we observed to the test time is short compared with the single scan path architecture. Because the ESP architecture uses the common bus, there are not additional signals in multi-board testing. By comparing the ESP architecture with conventional one using ISCAS '85 bench mark circuit, we showed that the architecture has improved results.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.1
• /
• pp.726-734
• /
• 2015

Relational databases used by structuralizing data are the most widely used in data management at present. However, in relational databases, service becomes slower as the amount of data increases because of constraints in the reading and writing operations to save or query data. Furthermore, when a new task is added, the database grows and, consequently, requires additional infrastructure, such as parallel configuration of hardware, CPU, memory, and network, to support smooth operation. In this paper, in order to improve the web information services that are slowing down due to increase of data in the relational databases, we implemented a model to extract a large amount of data quickly and safely for users by processing Hadoop Distributed File System (HDFS) files after sending data to HDFSs and unifying and reconstructing the data. We implemented our model in a Web-based civil affairs system that stores image files, which is irregular data processing. Our proposed system's data processing was found to be 0.4 sec faster than that of a relational database system. Thus, we found that it is possible to support Web information services with a Hadoop-based big data processing technique in order to process a large amount of data, as in conventional relational databases. Furthermore, since Hadoop is open source, our model has the advantage of reducing software costs. The proposed system is expected to be used as a model for Web services that provide fast information processing for organizations that require efficient processing of big data because of the increase in the size of conventional relational databases.

• Journal of KIISE:Information Networking
• /
• v.34 no.3
• /
• pp.226-238
• /
• 2007

In this paper, we present a novel fully distributed topology control protocol that can construct the multiple-ring topology of Minimal Connected Dominating Set (MCDS) as the transport backbone for mobile ad hoc networks. It makes a topology from the minimal nodes that are chosen from all the nodes, and the constructed topology is comprised of the minimal physical links while preserving connectivity. This topology reduces the interference. The all nodes work as the nodes of the distributed parallel Boltzmann machine, of which the objective function is consisted of two Boltzmann factors: the link degree and the connection domination degree. To define these Boltzmann factors, we extend the Connected Dominating Set into a fuzzy set, and also define the fuzzy set of nodes by which the multiple-ring topology can be constructed. To construct the transport backbone of the mobile ad hoc network, the proposed protocol chooses the nodes that are the strong members of these two fuzzy sets as the clusterheads. We also ran simulations to provide the quantitative comparison against the related works in terms of the packet loss rate and the energy consumption rate. As a result, we show that the network that is constructed by the proposed protocol has far better than the other ones with respect to the packet loss rate and the energy consumption rate.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.5A
• /
• pp.432-440
• /
• 2011

For the 4th generation mobile communication, LTE-advanced system needs the broad frequency band up to 100MHz for providing the data rate of maximum 1Gpbs. However, it is very difficult to secure the broad frequency band in the current frequency allocation situation. So, carrier aggregation was proposed as the solution, in which several fragmented frequency bands are used at the same time. Basically, multiple parallel receivers are required to get the information data from the different frequency bands but this conventional multi-chain receiver system is very inefficient. Therefore, in this paper, we like to study the single chain system that is able to receive the multi-band signals in a single receiver based on the time division multiplexing (TDM) reception method. This proposed TDM receiver efficiently manage to receive the multi-band signals in time domain and handle the baseband signals with one DSP board. However, the serious distortion could be generated by the sampling timing offset (STO) in the TDM-based system. Therefore, we like to analyze STO effects in the TDM-based system and propose a compensation method using estimated STO. Finally, it is shown by simulation that the proposed method is appropriate for the single chain receiver and show good compensation performance.

• Journal of KIISE:Software and Applications
• /
• v.36 no.4
• /
• pp.253-260
• /
• 2009

Many algorithms for computer vision and pattern recognition have recently been implemented on GPU (graphic processing unit) for faster computational times. However, the implementation has two problems. First, the programmer should master the fundamentals of the graphics shading languages that require the prior knowledge on computer graphics. Second, in a job that needs much cooperation between CPU and GPU, which is usual in image processing and pattern recognition contrary to the graphic area, CPU should generate raw feature data for GPU processing as much as possible to effectively utilize GPU performance. This paper proposes more quick and efficient implementation of neural networks on both GPU and multi-core CPU. We use CUDA (compute unified device architecture) that can be easily programmed due to its simple C language-like style instead of GPU to solve the first problem. Moreover, OpenMP (Open Multi-Processing) is used to concurrently process multiple data with single instruction on multi-core CPU, which results in effectively utilizing the memories of GPU. In the experiments, we implemented neural networks-based text extraction system using the proposed architecture, and the computational times showed about 15 times faster than implementation on only GPU without OpenMP.