• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.6B
• /
• pp.367-374
• /
• 2007

In this paper, we investigate the features of scientific remote collaborative applications. The presented features are the high speed disk access, the high speed transportation protocol, and the dedicated light path, which are for the high performance of remote end-to-end collaboration. In the environment that satisfies the features, we achieve 4.7Gbps peak throughput between Korea and U.S.A.

• Journal of KIISE:Information Networking
• /
• v.35 no.5
• /
• pp.415-424
• /
• 2008

In this paper, we compare the QoS performance enhancement schemes according to the network architecture of mobile WiMAX system in order to improve the end-to-end QoS performance and propose QoS parameter mapping method and IP packet scheduling algorithm. To evaluate the end-to-end QoS performance, we implemented an end-to-end simulator of mobile WiMAX system using OPNET. Simulation results show that the proposed QoS parameter mapping scheme reduces the average delay of VoIP packet and improves uplink resource efficiency. And also, when the proposed IP packet scheduling algorithm is applied to the system, the end-to-end packet transmission delay of VoIP service can be reduced by 44-67 percent compared to FIFO and WRR scheduler.

• Phonetics and Speech Sciences
• /
• v.12 no.2
• /
• pp.29-37
• /
• 2020

Recently, the neural network-based deep learning algorithm has dramatically improved performance compared to the classical Gaussian mixture model based hidden Markov model (GMM-HMM) automatic speech recognition (ASR) system. In addition, researches on end-to-end (E2E) speech recognition systems integrating language modeling and decoding processes have been actively conducted to better utilize the advantages of deep learning techniques. In general, E2E ASR systems consist of multiple layers of encoder-decoder structure with attention. Therefore, E2E ASR systems require data with a large amount of speech-text paired data in order to achieve good performance. Obtaining speech-text paired data requires a lot of human labor and time, and is a high barrier to building E2E ASR system. Therefore, there are previous studies that improve the performance of E2E ASR system using relatively small amount of speech-text paired data, but most studies have been conducted by using only speech-only data or text-only data. In this study, we proposed a semi-supervised training method that enables E2E ASR system to perform well in corpus in different domains by using both speech or text only data. The proposed method works effectively by adapting to different domains, showing good performance in the target domain and not degrading much in the source domain.

• Journal of KIISE:Information Networking
• /
• v.27 no.2
• /
• pp.160-174
• /
• 2000

Nowadays, most widely used transport protocol, TCP is tuned to perform well in traditional networks where packet losses occur mostly because of congestion. TCP performs reliable end-to-end packet transmission under the assumption of low packet error rate. However, networks with wireless links suffer from significant losses due to high error rate and handoffs. TCP responds to all losses by invoking congestion control and avoidance algorithms, resulting in inefficient use of network bandwidth and degraded end-to-end performance in that system. To solve this problem, several methods have been proposed. In this paper, we analyse and compare these methods and propose appropriate model for improving TCP performance in the network with wireless links. This model uses TCP selective acknowledgement (SACK) option between TCP ends, and also uses caching method at the base station. Our simulation results show that using TCP SACK option with base station caching significantly reduces unnecessary duplicate retransmissions and recover packet losses effectively.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.40 no.10
• /
• pp.35-42
• /
• 2003

In this paper, we classified the wavelength reservation mechanisms of optical burst switching (OBS), and investigated the path length priority effect of the offset time/delayed reservation (OT/DR)-based scheme like the just-enough-time (JET) protocol. And, we showed that the fiber delay line (FDL)-based scheme can avoid the path length priority effect in multiple hop network environments. Also, we proposed a novel FDL/DR with hop-by-hop priority increasing (HPI) scheme which can improve the end-to-end throughput by increasing the priority of bursts hop-by-hop using the input FDLs at each node. The simulation results showed that the proposed FDL/DR with HPI scheme could avoid the path length priority effect and enhance the end-to-end throughput in multiple hop network environments.

• Journal of KIISE:Information Networking
• /
• v.36 no.3
• /
• pp.184-190
• /
• 2009

WiBro has advantages when both WLAN and 3G UMTS are adopted. Much research is being carried out in this area. However, the WiBro specification does not consider end-to-end security. Hence, another security protocol has to be adopted to support secure communication. Most previous research only focused on WiBro MAC performance improvement or security. In this paper, we adopt a security protocol such as IPsec, TLS, and DTLS, well known end-to-end security protocols, to make full use of WiBro in the IP network. Using NS-2 we simulated the adoption of end-to-end security protocol and evaluated performance and usability. Simulation shows DTLS had some performance advantages. All the protocols, TLS and IPsec are also suitable for use in WiBro.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.151-154
• /
• 2020

뉴럴넷 기술이 발전과 힘께 다양한 분야에서 획기적인 성능 향상이 이루어지고 있다. 이미지 압축 분야에서도 기존의 전통적인 툴 제인 구조의 압축 방식에서 벗어나 종단간(end-to-end) 뉴렬넷 기반의 이미지 압축 기술에 대한 연구가 활발히 이루어지고 있다. 특히 최근 네트워크를 통해 변환된 피쳐 데이터의 엔트로피를 최소화하는 방식에 대한 연구가 활발히 이루어지고 있으며, 이에 기반한 최근의 연구는 VVC 화면 내 코딩 기술보다 우수한 코딩 효율성을 제공하고 있다. 그러나 변환된 피쳐 데이터에 대한 특성 분석은 부족한 실정이며, 이에 본 논문에서는 엔트로피 최소화 기반 종단간 이미지 압축 네트워크의 피쳐 공간 데이터에 대한 공간적 (spatial) 상관관계와 채널간(inter-channel) 상관관계(correlation)를 분석하고, 나아가 최근 제안된 종단간 이미지 압축 네트워크의 문맥 기반 예측 기능을 통해 잔존하는 데이터 중복성이 효과적으로 제거됨을 보인다.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.6
• /
• pp.578-586
• /
• 2021

We propose a semi-supervised learning method based on Variational AutoEncoder (VAE) and Unsupervised Data Augmentation (UDA) to improve the performance of an end-to-end speech recognizer. In the proposed method, first, the VAE-based augmentation model and the baseline end-to-end speech recognizer are trained using the original speech data. Then, the baseline end-to-end speech recognizer is trained again using data augmented from the learned augmentation model. Finally, the learned augmentation model and end-to-end speech recognizer are re-learned using the UDA-based semi-supervised learning method. As a result of the computer simulation, the augmentation model is shown to improve the Word Error Rate (WER) of the baseline end-to-end speech recognizer, and further improve its performance by combining it with the UDA-based learning method.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2003.05b
• /
• pp.1341-1344
• /
• 2003

최근 다양한 정보통신 서비스 창출과 향상된 성능을 가진 장치 개발 등으로 인해 정보통신서비스 시장이 확대되고 있다. 본 논문은 통신망에 대한 종단간 사용자의 서비스 질을 향상시키기 위하여, 인터넷의 대표적인 전달망인 IP 망에서의 QoS(Quality of Service)와 하부 전달망(ATM 망 프레임 릴레이망)에서의 QoS 간의 상관 관계를 손실 지표를 이용하여 분석함으로써 품질 지표간 맵핑 방안을 제안한다. 본 연구를 통해 종단간 서비스 품질 수준을 제공하기 위한 하부망에서의 손실의 수준 값을 알 수 있게된다.

• Journal of KIISE:Information Networking
• /
• v.31 no.6
• /
• pp.608-619
• /
• 2004

Application layer multicast(ALM) has been developed as an alternative to support the multi-receiver applications due to the slow deployment of IP multicast. Unlike routers in IP multicast, end systems participating in a ALM session are not optimized for relaying data and have various processing and network capacity. Therefore, with the increased role of end systems, the delay incurred at the end systems occupy large portion of the end-to-end delay. In this paper, we model the end system delay of ALM sessions and analyze its impact on the performance of the existing ALM schemes. Furthermore, we propose an enhanced scheme, based on the existing mesh-based scheme, which reflects the end system delay We also evaluate the performance of the proposed scheme by simulation and show that it reduces the end-to-end delay compared with the existing scheme.