• Journal of Korea Society of Digital Industry and Information Management
• /
• v.18 no.2
• /
• pp.17-26
• /
• 2022

In this paper, we propose a novel method of performing convolutional operations on a 2-D Processing Element(PE) array. The conventional method [1] of mapping the convolutional operation using the 2-D PE array lacks flexibility and provides low utilization of PEs. However, by mapping a convolutional operation from a 2-D PE array to a 1-D PE array, the proposed method can increase the number and utilization of active PEs. Consequently, the throughput of the proposed Deep Convolutional Neural Network(DCNN) accelerator can be increased significantly. Furthermore, the power consumption for the transmission of weights between PEs can be saved. Based on the simulation results, the performance of the proposed method provides approximately 4.55%, 13.7%, and 2.27% throughput gains for each of the convolutional layers of AlexNet, VGG16, and ResNet50 using the DCNN accelerator with a (weights size) x (output data size) 2-D PE array compared to the conventional method. Additionally the proposed method provides approximately 63.21%, 52.46%, and 39.23% power savings.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.671-672
• /
• 2011

비 신뢰적이고 비 대칭적인 무선 링크를 갖는 무선센서네트워크에서 데이터전송이 성공했음에도 ACK 메시지 전송 실패로 발생하는 불필요한 데이터 재전송을 피해 에너지 효율을 높이는 전송기법들이 제안되었다. 이러한 전송기법의 하나인 Probability-based Data Forwarding(PDF)기법에서 기대 전송 횟수 임계치를 설정해 임계값만큼 데이터를 재전송 하였다면 더 이상 재전송하지 않음으로 불필요한 에너지 낭비를 줄인다. 하지만 PDF 는 데이터 전송 횟수만을 고려함으로, 여전히 ACK 메시지의 신뢰성 있는 전송을 보장하지 못한다. 따라서 본 논문은 에너지 효율적인 PDF 기법에 비 대칭적 무선 노드의 특성을 고려하여 ACK 메시지 전송 시 사용되는 역방향 링크의 신뢰성이 낮다면 높은 신뢰성의 역방향 링크를 갖는 노드들을 선택하여 멀티 홉으로 송신 노드에게 ACK 메시지를 전송하는 기법을 적용한다. 이를 통해 불필요한 데이터 전송을 줄여 에너지의 낭비를 줄이고 무선센서네트워크의 수명을 연장한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.774-776
• /
• 2024

The effective implementation of advanced speech recognition (ASR) systems necessitates the deployment of sophisticated keyword spotting models that are both responsive and resource-efficient. The initial local detection of user interactions is crucial as it allows for the selective transmission of audio data to cloud services, thereby reducing operational costs and mitigating privacy risks associated with continuous data streaming. In this paper, we address these needs and propose utilizing the Whisper-Tiny model with fine-tuning process to specifically recognize keywords from google speech dataset which includes 65000 audio clips of keyword commands. By adapting the model's encoder and appending a lightweight classification head, we ensure that it operates within the limited resource constraints of local devices. The proposed model achieves the notable test accuracy of 92.94%. This architecture demonstrates the efficiency as on-device model with stringent resources leading to enhanced accessibility in everyday speech recognition applications.

• Journal of Advanced Navigation Technology
• /
• v.27 no.5
• /
• pp.610-620
• /
• 2023

The interest in the aircraft interior market is gradually growing due to technological development based on the 4th industrial revolution and competition for airlines to attract customers, and as part of that, Cabin Display System (CDS) for FAA Part.25 civil aircraft is being developed in Korea. The CDS is a system that provides various multimedia services to passengers by utilizing Flexible and Transparent Organic Light Emitting Diodes (OLED) with Integrated Display Processing Module (IDPM). This paper presents a new method for efficient Data Load Process of large-sized files and deals with their implementation and performance. The results of this study are expected to be applied to Data Load Process development of avionics that require reliable large-capacity file transmission along with reducing the costs of development compared to existing ARINC-615A.

• The KIPS Transactions:PartC
• /
• v.10C no.7
• /
• pp.923-928
• /
• 2003

This paper proposes a bottleneck link bandwidth measurement algorithm for reducing packet transmission delay within the grid network. There are two methods for measuring bottleneck link bandwidth：Packet Pair algorithm and Paced Probes algorithm. They measure bottleneck link bandwidth using the difference in arrival times of two paced probe packets of the same size traveling from the same source to destination. In addition, they reduce the influences of cross traffic by pacer packet. But there are some problems on these algorithms：it's not possible to know where bottleneck link occurred because they only focus on measuring the smallest link bandwidth along the path without considering bandwidth of every link on the path. So hop-by-hop based bottleneck link bandwidth measurement algorithm can be used for reducing packet transmission delay on grid network. Timestamp option was used on the paced probe packet for the link level measurement of bottleneck bandwidth. And the reducing of packet transmission delay was simulated by the solving a bottleneck link. The algorithm suggested in this paper can contribute to data transmission ensuring FTP and realtime QoS by detecting bandwidth and the location where bottleneck link occurred.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.4
• /
• pp.713-720
• /
• 2018

The Signal Intelligence (SIGINT) system based on Angle-of-Arrival(AOA) estimation, interference suppression, and transmission beamforming techniques is a cutting edge technology for efficiently collecting various signal information. In this paper, we present the efficient structure of a satellite system consisted of an AOA estimator, an adaptive beamformer, a signal processing and D/B unit, and a transmission beamformer, for collecting signal information. For accurately estimating AOAs of various signals, efficiently suppressing interference or jamming signals, and efficiently transmitting the collected information or data, we employ Multiple Signal Classification (MUSIC), Minimum Variance Distortionless Response (MVDR), and Minimum Mean Square Error (MMSE) algorithms, respectively. Also, we evaluate and analysis the performance of the presented satellite system through the computer simulation.

• The KIPS Transactions:PartC
• /
• v.9C no.6
• /
• pp.875-882
• /
• 2002

The active packets unlike traditional IP packets should be executed at each active node along their path. To execute the active program at each active node, the active packets for an active program should go through the same active nodes and all packets are delivered without any loss. This paper presents the new active engine for every active packet which execute an active program to be routed through the same intermediate active nodes and to be delivered reliably. Proposed active engine requires fewer changes to existing IP protocols and guarantees the reliable delivery of the active packets. Moreover, even if the sender does not have any information about the intermediate nodes every active packet is routed through the same intermediate active nodes. The simulation results show that proposed active engine achieves an efficient transmission with high data delivery and low communication overhead compared with the other existing transmission protocols.

• Journal of Electrical Engineering and Technology
• /
• v.10 no.6
• /
• pp.2384-2392
• /
• 2015

The technical development and practical applications of big-data for health is one hot topic under the banner of big-data. Big-data medical image fusion is one of key problems. A new fusion approach with coding based on Spherical Coordinate Domain (SCD) in Wireless Sensor Network (WSN) for big-data medical image is proposed in this paper. In this approach, the three high-frequency coefficients in wavelet domain of medical image are pre-processed. This pre-processing strategy can reduce the redundant ratio of big-data medical image. Firstly, the high-frequency coefficients are transformed to the spherical coordinate domain to reduce the correlation in the same scale. Then, a multi-scale model product (MSMP) is used to control the shrinkage function so as to make the small wavelet coefficients and some noise removed. The high-frequency parts in spherical coordinate domain are coded by improved SPIHT algorithm. Finally, based on the multi-scale edge of medical image, it can be fused and reconstructed. Experimental results indicate the novel approach is effective and very useful for transmission of big-data medical image(especially, in the wireless environment).

• International Journal of Precision Engineering and Manufacturing
• /
• v.9 no.4
• /
• pp.79-82
• /
• 2008

A method is described for ideally reconstructing the profile from a surface profiling measurement containing a reasonable amount of null measurement data. The proposed method can conjecture lost information and rectify irregular data that result due to bad measuring environments, signal transmission noise, or instrument-induced errors, The method adopts the concept of computer graphics and consists of several processing steps. First, a search for valid data in the neighborhood of the null data is performed. The valid data are then grouped and their contours are extracted. By analyzing these contours, a bounding box can be obtained and the general distribution of the entire area encompassing the valid and null data is determined Finally, an ideal surface model is overlaid onto the measurement results based on the bounding box, generating a complete reconstruction of the calculations, A surface-profiling task on a liquid crystal display photo spacer is used to verify the proposed method. The results are compared to those obtained through the use of a scanning electron microscope to demonstrate the accuracy of the proposed method.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.6
• /
• pp.10-14
• /
• 2015

In holographic data storage systems, data is recorded and read by page on a volume of storage medium, and it can increase transmission rate and storage capacity because of two-dimensional page-oriented data processing by charge-coupled devices. However, HDS suffers two-dimensional intersymbol interference unlike conventional data storages. In this paper, we propose a preprocessing method of decreasing ISI before read data in HDS pass to detector. This method has some advantage when we collaborate with the preprocessing for reducing misalignment error and modulation code.