• Proceedings of the KIEE Conference
• /
• 2003.11b
• /
• pp.3-6
• /
• 2003

뱀형 로봇은 일반적인 바퀴형 이동로봇과 운동 메카니즘이 상이하며 다관절로 이루어져 있기 때문에 장애물 회피에 있어 빠른 정보의 처리와 이를 위한 특별한 정보가 요구된다. 이를 실현하기 위하여 로봇은 자신의 위치를 지속적으로 파악하면서 장애물의 좌표 값과 일정한 거리의 간격을 두고 움직여야 한다. 주행 궤도 및 장애물 회피를 위한 알고리즘을 검증하기 위하여 가상 뱀형 시뮬레이터를 제작하였다. 시뮬레이터는 이동 주행 궤도를 생성하고, 지나온 궤도를 재현할 수 있는 재현기(Back Tracker), 앞으로 이루어질 뱀형 로봇의 위치와 자세를 알아보는 예견기(Predictor)로 구성된다. 시뮬레이터를 통하여 주위의 장애물을 안전하게 통과할 수 있는 일반적인 알고리즘인 포텐셜함수의 특성을 알아보고, 국소 최소점(Local Minima)에 빠지기 쉬운 단점을 극복하기 위한 방안을 제시한다. 본 논문에서는 뱀의 이동 주행 궤적을 알아보고, 주위의 장애물을 안전하게 통과할 수 있도록 하는 알고리즘에 대한 고찰과 제안한 알고리즘을 소프트웨어적인 3D 시뮬레이션을 통하여 걸과를 분석하고 검증한다.

• Annual Conference of KIPS
• /
• 2019.10a
• /
• pp.1101-1104
• /
• 2019

키넥트는 RGBD 카메라로 인체의 뼈대와 관절을 3D 공간에서 스켈레톤 데이터수집을 가능하게 해주었다. 스켈레톤 데이터를 활용한 행동 분류는 RNN, CNN 등 다양한 인공 신경망으로 접근하고 있다. 본 연구는 키넥트를 이용해서 Skeleton Joints를 수집하고, DNN 기반 스켈레톤 모델링 학습으로 행동을 분류한다. Skeleton Joints Processing 과정은 키넥트의 Depth Map 기반의 Skeleton Tracker로 25가지 Skeleton Joints 좌표를 얻고, 학습을 위한 전처리 과정으로 각 좌표를 상대좌표로 변경하고 데이터 수를 제한하며, Joint가 트래킹 되지 않은 부분에 대한 예외 처리를 수행한다. 스켈레톤 모델링 학습 과정에선 3계층의 DNN 신경망을 구축하고, softmax_cross_entropy 함수로 Skeleton Joints를 집는 모션, 내려놓는 모션, 팔짱 낀 모션, 얼굴을 가까이 가져가는 모션 해서 4가지 행동으로 분류한다.

• Journal of Information Display
• /
• v.11 no.3
• /
• pp.100-108
• /
• 2010

A multi-user autostereoscopic display based on laser scanning is described in this paper. It does not require the wearing of special glasses; it can provide 3D to several viewers who have a large degree of freedom of movement; and it requires the display of only a minimum amount of information. The display operates by providing regions in the viewing field, referred to as "exit pupils," which follow the positions of the viewers' eyes under the control of a multi-user head tracker. The display incorporates an RGB laser illumination source that illuminates a light engine. The light directions are controlled by a spatial light modulator, and a front screen assembly incorporates a novel Gabor superlens. Its operating principle is explained in this paper, as is the construction of three iterations of the display. Finally, a method of developing the display into one that is suitable for television applications is described.

• Journal of Broadcast Engineering
• /
• v.14 no.4
• /
• pp.419-427
• /
• 2009

We present a novel head tracking system for stereoscopic displays that ensures the viewer has a high degree of movement. The tracker is capable of segmenting the viewer from background objects using their relative distance. A depth camera using TOF(Time-Of-Flight) is used to generate a key signal for eye tracking application. A method of the moving parallax barrier is also introduced to supplement a disadvantage of the fixed parallax barrier that provides observation at the specific locations.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.17 no.10
• /
• pp.907-911
• /
• 2007

Compliance of joints must be considered when we analyze dynamics of a multi-body system. If the virtual model for CAE(computer aided engineering) analysis does not consider compliance, the result of CAE analysis can be very different from the actual experimental result. Especially in a biped walking robot, the robot may lose walking stability due to the compliance in joints of a walking robot. This paper proposed a method applying a compliance of joints in the biped walking robot to a virtual model. Also, through the 3-D displacement measurement using a laser tracker, it was demonstrated that the virtual model considering the joint compliance could effectively simulate the nonlinear motion of the real model.

• Solar Energy
• /
• v.18 no.3
• /
• pp.129-135
• /
• 1998

A photovoltaic system is an infinite and clean energy system. A photovoltaic system consists of a solar cell array, a converter, a inverter and a control unit. It is necessary that the Maximum Power Point Tracker(MPPT) is applied to the photovoltaic system because the output power of solar cell array is varied with irradiation, temperature and external effects. In this paper, the neural networks theory, one of the control methods, is applied to track the maximum power point of the photovoltaic system. The MPPT using neural networks theory is proposed to improve existing energy converter efficiency. Also the theory is applied to operation of inverter type airconditionig system.

• The Journal of the Korea Contents Association
• /
• v.16 no.9
• /
• pp.451-459
• /
• 2016

While an object detection algorithm plays a key role in many computer vision applications, it requires extensive computation to show robustness under varying lightning and geometrical distortions. Recently, some approaches formulate the problem in a classification framework and show improved performances in object recognition. Among them, random fern algorithm drew a lot of attention because of its simple structure and high recognition rates. However, it reveals performance degradation under the illumination changes and noise addition, since it computes patch features based only on pixel intensities. In this paper, we propose a new structure of combined random fern which incorporates depth information into the conventional random fern reflecting 3D structure of the patch. In addition, a new structure of object tracker which exploits the combined random fern is also introduced. Experiments show that the proposed method provides superior performance of object detection under illumination change and noisy condition compared to the conventional methods.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.1B
• /
• pp.24-33
• /
• 2002

The cell search scheme for W-CDMA consists of the following three stages: slot synchronization(1st stage), group identification and frame boundary detection(2nd stage), and long code identification(3rd stage). The performance of the cell search when a mobile station is switched on, which is referred to as initial cell search, is decreased by the initial frequency and timing error. In this paper, we propose the pipeline structured initial cell search scheme using time trackers to compensate for the impact of the initial timing error in the stage 2 and stage 3. The simulation results show that the performance of the proposed scheme is maximal 1.5dB better than that of the conventional one when the initial timing error is near ${\pm}T_c$/2.

• Proceedings of the KSME Conference
• /
• 2008.11a
• /
• pp.1709-1712
• /
• 2008

The trunk is inclined to the loaded side when carrying an object as one of activities of daily living. As the reaction to this behavior the human body may be inclined to his/her trunk to unloaded side. The present study investigated the biomechanical effects of weight variation for sided load carriage during walking upon joint moments and muscle torques, through the tracker agent and joint driving dynamic analysis. To perform the experiment one male was selected as subject for the study. Gait analysis was performed by using a 3D motion analysis system. Thirty nine 14mm reflective markers, according to the plug-in marker set, were attached to the subject. We used BRG.LifeMOD(Biomechanics Research Group, Inc., USA), for skeletal modeling and inverse and joint driving dynamic simulation during one gait cycle. In walking with a sided load carriage, the subject modeled held the carriage with the right hand, which weighed 0, 5, 10, 15kg, 20kg respectively. The result of this simulation showed that knee and hip in the coronal plane were inclined to the loaded side and loaded side had larger moments as the sided load carriage was increased. On the other hand thoracic and lumbar in the coronal plane had larger negative values as the sided loaded carriage was increased. The thoracic and lumbar in the transverse plane also had larger values as the sided load was increased. And the several muscles of loaded side were increased as increasing sided load. It could be concluded that human body is adopted to side loaded circumstances by showing more biologic force. These results could be very useful in analysis for delivery motion of daily life.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.05a
• /
• pp.877-881
• /
• 2008

In this paper, we propose a real-time people tracking system with multiple CCD cameras for security inside the building. The camera is mounted from the ceiling of the laboratory so that the image data of the passing people are fully overlapped. The implemented system recognizes people movement along various directions. To track people even when their images are partially overlapped, the proposed system estimates and tracks a bounding box enclosing each person in the tracking region. The approximated convex hull of each individual in the tracking area is obtained to provide more accurate tracking information. To achieve this goal, we propose a method for 3D walking human tracking based on the IMPRESARIO framework incorporating cascaded classifiers into hypothesis evaluation. The efficiency of adaptive selection of cascaded classifiers have been also presented. We have shown the improvement of reliability for likelihood calculation by using cascaded classifiers. Experimental results show that the proposed method can smoothly and effectively detect and track walking humans through environments such as dense forests.