• Journal of the Korea Institute of Military Science and Technology
• /
• v.12 no.5
• /
• pp.563-570
• /
• 2009

In the future, most military activities will be replaced by robots. Because of many dangerous factors in battlefield, reconnaissance should be performed by robot. Reconnaissance robot should be small for not being detected, be light and simple structure for personal portability and overcome unexpected rough terrain for mission completion. In case of small and light robot, it can't get enough friction force for movement. Therefore small reconnaissance robot need jumping function for movement. In this paper we proposed a biologically inspired jumping mechanism. And we adjusted moment and jumping angle by using four bar linkage, especially varying coupler length.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2013.05a
• /
• pp.57-58
• /
• 2013

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.311-314
• /
• 2004

가장 완벽한 지능형 모델로 알려져 있는 두뇌는 인공 지능을 구현하기 위해 이해되어야 하는 많은 내용을 지니고 있다. 하지만, 현재까지는 두뇌의 생물학적인 정보처리 메커니즘은 극히 일부분에서 밝혀졌고 대부분의 내용은 추측이나 가정으로 설명되고 있다. 이미 밝혀진 두뇌의 정보처리 메커니즘에 기반한 정보처리 시스템은 다양한 응용 분야에 활용되어 지금의 시스템보다 월등한 성능을 보일 것으로 예상된다. 이에, 본 논문에서는 두뇌의 생물학적 흐름을 카테고리 별로 정리하였으며 이를 구현할 수 있는 소프트 컴퓨팅 기법을 소개한다. 다양한 소프트 컴퓨팅 기법을 이용하여 구현된 인공 두뇌 모델은 정보처리 과정에서 자율적이며, 효과적인 정보처리 성능을 보여줌을 알 수 있다. 이는 인공 지능 시스템의 새로운 도약에 필요한, 정형화된 모델로 활용될 수 있을 것으로 기대된다.

• Proceedings of the KSME Conference
• /
• 2007.05a
• /
• pp.1195-1200
• /
• 2007

It is desirable that the guard robot should be small-sized and light-weighted to increase its portability. In addition, it should be able to overcome a relatively high obstacle to cope with different situations. The jumping robot can reach a higher place more rapidly than other locomotion methods. This research proposes the jumping mechanism based on the conical spring for a small robot. Both the clutch mechanism and conical spring are incorporated into the jumping mechanism. In the clutch mechanism, the spring can be immediately compressed and released by one actuator with the planetary gear train and one-way clutch. The robot equipped with the jumping mechanism can overcome the obstacles which are higher than its height. In this paper, the characteristic of the conical spring for the jumping robot is determined and the small-sized, lightweight jumping mechanism is developed. The validity of the jumping mechanism was verified by various experiments. It is shown that the robot using this mechanism can provide good mobility in the rough terrain.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.11a
• /
• pp.377-380
• /
• 2005

가장 완벽한 지능형 모델로 알려져 있는 두뇌는 인공 지능을 구현하기 위해 이해되어야 하는 많은 내용을 지니고 있다. 하지만, 현재까지는 두뇌의 생물학적인 정보처리 메커니즘은 극히 일부분에서 밝혀졌고 대부분의 내용은 추측이나 가정으로 설명되고 있다. 이미 밝혀진 두뇌의 정보처리 메커니즘에 기반한 정보처리 시스템은 다양한 응용 분야에 활용되어 지금의 시스템보다 월등한 성능을 보일 것으로 예상된다. 이에, 본 논문에서는 두뇌의 생물학적 흐름을 카테고리 별로 정리하였으며 이를 구현할 수 있는 소프트 컴퓨팅 기법을 소개한다. 다양한 소프트 컴퓨팅 기법을 이용하여 구현된 인공두뇌 모델은 정보처리 과정에서 자율적이며, 효과적인 정보처리 성능을 보여줌을 알 수 있다. 이는 인공 지능 시스템의 새로운 도약에 필요한, 정형화된 모델로 활용될 수 있을 것으로 기대 된다.

• Journal of the Korean Magnetics Society
• /
• v.15 no.2
• /
• pp.81-84
• /
• 2005

The origin for the charge disproportionation (CD) transition in polycrystalline $La_{1/3}Sr_{2/3}FeO_{2.96}$ was examined using X-ray diffraction and the external field $M\ddot{o}ssbauer$ssbauer spectroscopy. In order to see how the external magnetic field affects the CD state above its transition temperature, an external magnetic field of up to 6 T was applied either parallel or perpendicular to the $\gamma-ray$ direction with the sample temperature fixed at 225 K, which was above the CD transition temperature. Without an external magnetic field, a completely paramagnetic singlet was obtained in the temperature range of the averaged valence state above the transition temperature, which was interpreted as coming from the average valence $Fe^{3.6+}$. In the longitudinal geometry, a magnetic Zeeman with its intensity ratio 3:0:1:1:0:3 is superimposed to the central singlet. In the transverse geometry, however, the central singlet disappears and only a magnetic component with its intensity ratio 3:4:1:1:4:3 emerges. The existence of a singlet is understood as an evidence of the fast electron-transfer among Fe ions. Since the singlet still exists under the magnetic field, the application of an external field has little effect on the conduction mechanism of hopping electrons.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.1
• /
• pp.24-31
• /
• 2020

In this paper, the deep convolutional neural network for sound event detection is studied. Especially, the end-to-end neural network, which generates the detection results from the input audio waveform, is studied for weakly supervised problem that includes weakly-labeled and unlabeled dataset. The proposed system is based on the network structure that consists of deeply-stacked 1-dimensional convolutional neural networks, and enhanced by the skip connection and gating mechanism. Additionally, the proposed system is enhanced by the sound event detection and post processings, and the training step using the mean-teacher model is added to deal with the weakly supervised data. The proposed system was evaluated by the Detection and Classification of Acoustic Scenes and Events (DCASE) 2019 Task 4 dataset, and the result shows that the proposed system has F₁-scores of 54 % (segment-based) and 32 % (event-based).