• 제어로봇시스템학회논문지
• /
• 제4권5호
• /
• pp.644-650
• /
• 1998

In this paper , a state space model for flexible beam is presented using the assumed-modes approach. The state space equation is derived for a flexible beam in which one end is connected to a motor and is driven by a torque equation and the other end is free. Many of the transfer function proposed thus far use the torque to the flexible beam as the input and the tip deflection of the flexible beam as the output. The Technique for the analysis and synthesis of the dual-input describing function(DIDF) is introduced here and the construction of a non-linear compensator, based on this technique, is proposed. This non-linear compensator, properly connected in the direct path of a closed-loop linear or non-linear control system. The above non-linear network is used to compensate linear and non-linear systems for instability, limit cycles, low speed of response and static accuracy. The effectiveness of the proposed scheme is demonstrated through computer simulation and experimental results.

• Bulletin of the Korean Chemical Society
• /
• 제18권5호
• /
• pp.501-509
• /
• 1997

In recent papers[Bull. Kor. Chem. Soc. 1996, 17, 735; ibid 1997, 18, 478] we reported results of molecular dynamics (MD) simulations for the thermodynamic, structural, and dynamic properties of liquid normal alkanes, from n-butane to n-heptadecane, using three different models. Two of the three classes of models are collapsed atomic models while the third class is an atomistically detailed model. In the present paper we present results of MD simulations for the corresponding properties of liquid branched-chain alkanes using the same models. The thermodynamic property reflects that the intermolecular interactions become weaker as the shape of the molecule tends to approach that of a sphere and the surface area decreases with branching. Not like observed in the straight-chain alkanes, the structural properties of model Ⅲ from the site-site radial distribution function, the distribution functions of the average end-to-end distance and the root-mean-squared radii of gyration are not much different from those of models Ⅰ and Ⅱ. The branching effect on the self diffusion of liquid alkanes is well predicted from our MD simulation results but not on the viscosity and thermal conductivity.

• International Journal of Internet, Broadcasting and Communication
• /
• 제13권1호
• /
• pp.210-218
• /
• 2021

CNN (Convolutional Neural Network), a type of deep learning algorithm, is a type of artificial neural network used to analyze visual images. In deep learning, it is classified as a deep neural network and is most commonly used for visual image analysis. Accordingly, an AI autonomous driving model was constructed through real-time image processing, and a crosswalk image of a road was used as an obstacle. In this paper, we proposed a low-cost model that can actually implement autonomous driving based on the CNN model. The most well-known deep neural network technique for autonomous driving is investigated and an end-to-end model is applied. In particular, it was shown that training and self-driving on a simulated road is possible through a practical approach to realizing lane detection and keeping.

• 방사성폐기물학회지
• /
• 제19권4호
• /
• pp.435-446
• /
• 2021

This article examines the consequences of a significant spent fuel management decision or event in the United States, South Korea and Taiwan. For the United States, it is the financial impact of the Department of Energy's inability to take possession of spent fuel from commercial nuclear power companies beginning in 1998 as directed by Congress. For South Korea, it is the potential financial and socioeconomic impact of the successful construction, licensing and operation of a low and intermediate level waste disposal facility on the siting of a spent fuel/high level waste repository. For Taiwan, it is the operational impact of the Kuosheng 1 reactor running out of space in its spent fuel pool. From these, it draws six broad lessons other countries new to, or preparing for, nuclear energy production might take from these experiences. These include conservative planning, treating the back-end of the fuel cycle holistically and building trust through a step-by-step approach to waste disposal.

• ETRI Journal
• /
• 제45권4호
• /
• pp.678-689
• /
• 2023

This paper describes a novel end-to-end deep generative model-based speaker recognition system using prosodic features. The usefulness of variational autoencoders (VAE) in learning the speaker-specific prosody representations for the speaker recognition task is examined herein for the first time. The speech signal is first automatically segmented into syllable-like units using vowel onset points (VOP) and energy valleys. Prosodic features, such as the dynamics of duration, energy, and fundamental frequency (F₀), are then extracted at the syllable level and used to train/adapt a speaker-dependent VAE from a universal VAE. The initial comparative studies on VAEs and traditional autoencoders (AE) suggest that the former can efficiently learn speaker representations. Investigations on the impact of gender information in speaker recognition also point out that gender-dependent impostor banks lead to higher accuracies. Finally, the evaluation on the NIST SRE 2010 dataset demonstrates the usefulness of the proposed approach for speaker recognition.

• 한국정보과학회논문지:정보통신
• /
• 제34권3호
• /
• pp.218-225
• /
• 2007

현재 홈 네트워크, 센서 네트워크, 유비쿼터스 네트워크 등에 대한 활발한 연구가 진행되면서 무선 이동 애드-혹 네트워크에 대한 관심이 높아지고 있다. 무선 이동 애드-혹 네트워크란 기존의 유선하부구조의 도움 없이 이동 호스트들만으로 구성되는 임시적인 네트워크로, 언제 어디서나 컴퓨팅 환경을 이용할 수 있도록 하는 개념의 유비쿼터스 컴퓨팅 환경에 적합한 네트워크이다. 본 논문에서는 다중 경로의 수를 제한하는 소스 라우팅 프로토콜을 기반으로 무선 이동 애드-혹 네트워크의 확장 기법을 제안한다. 이 기법은 무선 이동 애드-혹 네트워크 내의 이동 호스트들 간 또는 무선 이동 애드-혹 네트워크 내의 이동 호스트와 유선 네트워크 서비스를 지원하는 기지국 간에 링크 및 중간 호스트의 중복을 허용하는 다중 경로를 유지함으로써 경로 재설정 및 재등록으로 인한 오버헤드를 줄이는 기법이다. 이와 같이 다중 경로를 유지함으로써 출발지와 목적지간 데이타 패킷 전송 지연 및 패킷 손실을 줄일 수 있으며, 따라서 보다 신뢰적인 방법으로 무선 이동 애드-혹 네트워크를 유선 네트워크로 확장시킬 수 있다. 성능 평가를 통해, 본 논문에서 제안하는 기법이 단일 경로를 사용하는 일반적인 확장 기법들보다 이동 속도 증가에 따른 처리량 및 단대단 지연이 보다 안정적이며, 노드/링크가 전혀 중복되지 않는 다중 경로를 사용하는 기법에 비해 낮은 오버헤드를 가짐을 보인다.

• 로봇학회논문지
• /
• 제19권1호
• /
• pp.117-129
• /
• 2024

This paper introduces collective navigation through a narrow gap using a curriculum-based deep reinforcement learning algorithm for a swarm of unmanned aerial vehicles (UAVs). Collective navigation in complex environments is essential for various applications such as search and rescue, environment monitoring and military tasks operations. Conventional methods, which are easily interpretable from an engineering perspective, divide the navigation tasks into mapping, planning, and control; however, they struggle with increased latency and unmodeled environmental factors. Recently, learning-based methods have addressed these problems by employing the end-to-end framework with neural networks. Nonetheless, most existing learning-based approaches face challenges in complex scenarios particularly for navigating through a narrow gap or when a leader or informed UAV is unavailable. Our approach uses the information of a certain number of nearest neighboring UAVs and incorporates a task-specific curriculum to reduce learning time and train a robust model. The effectiveness of the proposed algorithm is verified through an ablation study and quantitative metrics. Simulation results demonstrate that our approach outperforms existing methods.

• 수산해양기술연구
• /
• 제33권1호
• /
• pp.59-67
• /
• 1997

This paper gives a controller design method by Linear Matrix Inequality(LMI) for internal combustion engine with Continuously Variable Transmission(CVT) which satisfies the given $H_\infty$ control performance and robust stability in the presence of physical parameter perturbations. To the end, the validity and applicability of this approach are illustrated by simulation in the all engine operating regions.

• 제어로봇시스템학회논문지
• /
• 제9권10호
• /
• pp.822-831
• /
• 2003

A computer-vision based pipe-inspection algorithm is developed. The algorithm uses the modified Hough transformation and a line-scanning approach to identify the edge line and the radius of the pipe image, from which the eccentricity and dimension of the pipe-end is calculated. Line and circle detection was performed using Laplace operator with input image, which are acquired from the front and side cameras. In order to minimize the memory usage and the processing time, a clustering method with the modified Hough transformation is introduced for line detection. The dimension of inner and outer radius of pipe is calculated by the proposed line-scanning method. The method scans several lines along the X and Y axes, calculating the eccentricity of inner and outer circle, by which pipes with wrong end-shape can be classified and removed.

• Journal of electromagnetic engineering and science
• /
• 제16권4호
• /
• pp.241-247
• /
• 2016

This paper proposes the design and measurement of a 6-18 GHz front-end receiver module that has been combined into a one- channel output from a two-channel input for electronic warfare support measures (ESM) applications. This module includes a limiter, high-pass filter (HPF), power combiner, equalizer and amplifier. This paper focuses on the design aspects of reducing the noise figure (NF) and matching the phase and amplitude. The NF, linear equalizer, power divider, and HPF were considered in the design. A broadband receiver based on a combined configuration used to obtain low NF. We verify that our receiver module improves the noise figure by as much as 0.78 dB over measured data with a maximum of 5.54 dB over a 6-18 GHz bandwidth; the difference value of phase matching is within $7^{\circ}$ between ports.