http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
A Practical Path Planner for the Robotic Vacuum Cleaner in Rectilinear Environments
Nakju Lett Doh,Chanki Kim,Wan Kyun Chung IEEE 2007 IEEE transactions on consumer electronics Vol.53 No.2
<P>In this paper, we propose a path planner for a robotic vacuum cleaner (RVC). In the design of the planner, we consider two main issues: (1) human-friendly path generation and (2) low computational load. First, we analyze how human move and suggest a hypothesis that human navigate in a way that minimizes the sum of muscle and brain energy. By imitating the human path, we propose a human- friendly path planner. Also, the designed planner requires a low amount of computations which not only extends the battery running time but also decreases the hardware cost of the RVC. Experimental results show that the proposed path gives a more favorable impression to customers than conventional paths.</P>
정낙주(NakJu Jeong),김주만(JooMan Kim),김해진(HaeJin Kim),임기욱(KiWook Rim) 한국정보과학회 1995 한국정보과학회 학술발표논문집 Vol.22 No.2B
본 논문은 주전산기 IV 로서 개발중인 고속병렬컴퓨터(SPAX)에 있어서 ION 운영체제를 위한 구조 및 설계를 제시한다. SPAX 는 입출력 성능을 높이기 위하여 프로그램을 수행하는 프로세싱 노드(PN)와 별도로 독립적인 운영체제를 탑재하고 있는 ION 을 따로 두고 있다. ION 운영체제는 기능에 따라 여러 개의 서버로 구성 된다. ION 서버는 블록 장치로부터의 입출력 성능을 높이기 위한 디스크 캐시 관리자, 블록 입출력 관리자, 고속으로 데이타를 백업하기 위한 고속 백업 관리자, CCN 의 콘솔에 이상이 생겼을 경우 콘솔의 기능을 대체하는 백업 콘솔 관리자를 갖도록 설계 하였다. ION 은 블록 입출력에 있어서 최대의 성능을 갖도록 서버들을 설계 하였으며, 입출력이 별도의 노드로 분리된 다중 노드 시스템에 유용하다.
Automatic CAD-structure extraction from 3D point clouds
Jihyeon Kwon,Nakju Lett Doh 한국정보통신학회 2015 2016 INTERNATIONAL CONFERENCE Vol.7 No.1
When we cannot use an original blueprint due to undocumented interior renovation or loss, we have measured and drawn structures manually. However, it is too time-consuming and inconvenient. Moreover, many objects placed in a room often occlude corners and edges of the room, which makes 3D indoor structure modeling more difficult. In this paper, we propose a new method for 3D indoor structure reconstruction which results in a simple CAD-structure with automatic process. For data collection of the indoor environment, we used a 3D laser sensor. For modeling structures, we classify the obtained 3D point clouds as structures and objects. Plane extraction among structure-PCD is followed since an extracted plane is intuitively part of a wall considering that indoor space is enclosed with walls. From the extracted planes, we seek points where three planes intersect since some intersection points are occluded by objects placed at the corner. Then several valid intersection points are determined as real structure points by the Plane-Patch test, eliminating other points that do not really exist. After that, we connect the valid structurepoints belonging to the ceiling, which creates a contour. With this, we can generate a floor plan of an indoor room. Considering the distance between the ceiling and the floor, finally, we reconstruct the 3D indoor structure as a simple CAD model. We applied this method to an indoor environment which consists of four rooms and obtained the results successfully.