• The Journal of the Acoustical Society of Korea
• /
• v.36 no.6
• /
• pp.419-424
• /
• 2017

TDoA (Time Difference-of Arrival) or DoA (Direction-of-Arrival) can be used for source localization. However, the localizing performance is dependent on relative position between source and receivers, receivers' geometric structure, sound speed, and so on. In this paper we propose a source localization method with enhanced performance that combines multiple information. The proposed method uses the time TDoA, DoA and sound speed as variables. LM (Levenberg-Marquardt) method which is one of nonlinear optimizations is applied. The performances of the proposed method was evaluated by simulation. As result of simulation, the proposed method has the lower average localizing error performance than the previous method.

• Journal of Korea Multimedia Society
• /
• v.12 no.10
• /
• pp.1436-1449
• /
• 2009

The value of sensing information is decided according to positions of sensor nodes, which are very important in sensor networks. In this paper, we propose a method that estimates positions of nodes by using adjacent node information and received signal strength in a sensor network. With the proposed method, we can find positions of nodes easily because we use information that nodes have. Moreover, we can find distribution easily for all the nodes because we can measure a relative position for a node whose position is not known based on anchor nodes whose positions are already known. We utilized Use case diagram, activity diagram and State machine diagram among several diagrams of UML to implement proposed method in sensor networks that is dynamic system. We can understand exact flow for each function of the proposed method in node position estimation system can be implemented easily. And we can be confirmed that the position of estimated nodes has a little error.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2012.07a
• /
• pp.91-94
• /
• 2012

본 논문에서는 3D 콘텐츠 제작에 있어 입체피로를 최소화하기 위한 영상의 수직시차 제거과정에 있어 사진측량기반의 강인기법을 제안하고자 한다. 영상의 수직시차 제거 과정은 크게 기하추정 단계와 에피폴라 변환 단계로 구성된다. 본 논문에서는 기하추정을 위해 사진측량에서 널리 활용되고 있는 공면조건 기반 상대표정 알고리즘을 적용한다. 이 때 상대표정 알고리즘에는 자동 정합점 추출에 따른 오정합과 위치오차에 강인성을 확보하기 위해 제약조건을 도입한 베이지안 접근방식을 적용하고자 하며, 이를 바탕으로 수행되는 에피폴라 변환에는 영상의 왜곡과 원 영상 대비 변형을 최소화하기 위해 공선조건기반의 중심투영변환기법을 적용하고자 한다. 알고리즘의 성능검증을 위한 비교 알고리즘으로, 기하추정에는 8점 알고리즘과 스테레오 캘리브레이션 기법이 사용되었으며 에피폴라 변환에는 Hartley 방법과 Bouguet 방법이 사용되었다. 실험결과는 제안알고리즘의 높은 정확도와 강인성, 그리고 활용적 가치를 보여주었다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.12
• /
• pp.1445-1452
• /
• 2013

In this paper, we propose a real-time algorithm for estimating the relative position of a person with respect to a robot (camera) using a monocular camera. The algorithm detects the head and shoulder regions of a person using HOG (Histogram of Oriented Gradient) feature vectors and an SVM (Support Vector Machine) classifier. The size and location of the detected area are used for calculating the relative distance and angle between the person and the camera on a robot. To increase the speed of the algorithm, we use a GPU and NVIDIA's CUDA library; the resulting algorithm speed is ~ 15 Hz. The accuracy of the algorithm is compared with the output of a SICK laser scanner.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.34 no.5
• /
• pp.74-80
• /
• 2006

GPS, which has been opened to the public since the 1980's, uses the C/A code time of arrival to estimate the position, and measures the carrier doppler frequency to estimate the velocity. In development from the 1990's, DGPS has improved position accuracy by eliminating common errors and CDGPS has achieved cm-level position accuracy using carrier phase. In this paper, a modified LSAST ambiguity resolution method for CDGPS is proposed to improve reliability and computational efficiency. Also the test results of cm level relative positioning of a moving vehicle using single frequency GPS receivers are compared to INS position. This research result can be widely used for the development of high precision INS, unmanned autonomous driving, survey and mapping, etc.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2014.04a
• /
• pp.872-875
• /
• 2014

재난방지 및 구호에 사용되는 로봇의 주된 목적은 인간이 직접적인 접근을 할 수 없는 지역을 사전에 탐사하여 인간으로 하여금 올바른 판단을 돕기 위함에 있다. 하지만, 재난 지역에서는 통신장애 문제나, 육안 식별이 불가능한 상황, 원격조정을 통하여 로봇이 업무 수행에 상당한 제약을 받는다. 이 문제를 해결하기 위해 "LED-RGB 칼라센서를 이용한 상호위치 인식 방법연구"[1]을 수행하였으나, RGB의 인식거리가 상당히 짧고, 판단이 모호한 단점이 발생하였다. 따라서 본 연구에서는 이를 개선한 RGB-D센서를 이용하여 RGB의 인식거리를 증가시켰다. 또한 더욱 높은 정확성을 이용하기 위해, Depth를 사용하여 사물들의 특징점들을 랜드마크로 하고 랜드마크로부터의 상대위치를 파악하여 위치를 추정하는 방법을 제안하고자 한다. 마지막으로 상호인식 알고리즘을 이전 방식과 비교하고자 한다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.2
• /
• pp.298-303
• /
• 2010

robots determine the location of the other robot using wireless sensors. Use it to decide how to move his. And go to any location, will make shape of column and line, circle. In this paper, we discuss problem in circle formation enclosing target which moves. It is method about enclosed invader in circle formation based on mutual localization of swarm robot without infrastructure. Therefore, use trilateration that do not need to know the value of the coordinates of reference points. So, Specify enclosed point for the number of robots base on between the relative position of the robot in the coordinate system. And particle filter is proposed to improve the accuracy of the location.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.3
• /
• pp.60-66
• /
• 2012

In this paper, we propose abandoned object detection algorithm. When abandoned object was occluded other object, the existing methods cannot detect abandoned object because those methods are not able to estimate the location of abandoned object. In order to overcome this problem, the proposed algorithm extracts the corners around abandoned object. The detected corners are linked to center of abandoned object called by supporters. We can then estimate the location of abandoned object by using supporters. Therefore, the proposed algorithm can detect and estimate the location of abandoned object, when abandoned object is occluded by other object. For this reason, the proposed algorithm can be applied to intelligent surveillance system to prevent bomb terror, which disguises as luggage or box.

• Spatial Information Research
• /
• v.14 no.4 s.39
• /
• pp.433-449
• /
• 2006

Two primary objectives of this study were to determine important personal factors in performing cognitive distance mapping, and to understand how human's cognitive distance mapping capabilities were influenced by reference and subjects' locations using Global Positioning System (GPS). Undergraduate and graduate students at the University of Kansas, USA were interviewed and surveyed throughout the campus area giving them a paper-and-pencil test. Study results showed that females had more accurate cognitive mapping capability than males regardless of ethnic background and academic levels. Generally, subjects with longer affiliation with the university, higher ages and academic levels had less variability in their mapping accuracy. Subjects tended to more accurately map the target locations closer to the reference points than those located farther away, and subjects who were closer to a reference point performed their distance mapping better than those farther away. A correlation analysis reported that male subjects used reference-to-target and subject-to-reference distances more sensitively than females to estimate the locations of the targets. This result indicates that males might have used the reference point-based map scale more strenuously than females.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10d
• /
• pp.592-594
• /
• 2002

얼굴 영역 검출의 수행 방법으로 개선된 얼굴 칼라 히스토그램과 에지 정보를 결합한 검출 시스템을 제안한다. 배경이 복잡한 영상에서 사람의 얼굴 영역과 배경 영역이 얼굴 영역과 비슷한 칼라 분포를 가지는 물체를 포함하는 영상이더라도 강인한 추출이 가능하도록 하였다. 본 논문에서는 효율적인 얼굴 검출을 위하여 얼굴의 칼라 분포를 얼굴 칼라의 확률 히스토그램으로 모델링하고 에지 정보와 reconstruction에 의한 형태학적 필터링(morphological filtering)을 적용하여 얼굴 후보 영역을 검출한다. 검출된 후보 영역에서 얼굴 구성 요소간의 위치 관계를 이용하여 눈동자와 흰자위의 명도차 특성으로 눈 영역의 위치를 추정하고 상대적인 위치 관계로 입 영역을 추정하여 얼굴 구성 요소의 정보를 얻어서이 요소 정보가 존재하는 후보 영역들이 최종적으로 얼굴 영역으로 판단되어 검출된다. 제안한 방법을 여러 영상에 이용하여 좋은 결과를 얻을 수 있었다.