• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 춘계학술대회 논문집
• /
• pp.1050-1053
• /
• 2005

There is no agreed definition of intelligence. The ability to adapt to the environments is a kind of intelligence. Expert functionally recognize environment using their five senses, and acquire and memorize knowledge necessary for operating machines. Knowledge that they cannot acquire directly is acquired in indirect ways. The purpose of intelligence machines is applying to machines experts' knowledge acquisition process and their skills in operating machine. An agent is an autonomous process that recognizes external environment, exchanges knowledge with external machines and performs an autonomous decision-making function in order to achieve common goals. This paper describes agent application for intelligence machine.

• Agricultural and Biosystems Engineering
• /
• 제6권2호
• /
• pp.47-53
• /
• 2005

AMachine vision is a promising tool for the autonomous guidance of farm machinery. Conventional CCD camera for the machine vision needs a desktop PC to install a frame grabber, however, a web camera is ready to use when plugged in the USB port. A web camera with a notebook PC can replace existing camera system. Autonomous steering control system of this research was intended to be used for combine harvester. If the web camera can recognize cut/uncut edge of crop, which will be the reference for steering control, then the position of the machine can be determined in terms of lateral offset and heading angle. In this research, a white line was used as a cut/uncut edge of crop for steering control. Image processing algorithm including capturing image in the web camera was developed to determine the desired travel path. An experimental vehicle was constructed to evaluate the system performance. Since the vehicle adopted differential drive steering mechanism, it is steered by the difference of rotation speed between left and right wheels. According to the position of vehicle, the steering algorithm was developed as well. Evaluation tests showed that the experimental vehicle could travel within an RMS error of 0.8cm along the desired path at the ground speed of $9\sim41cm/s$. Even when the vehicle started with initial offsets or tilted heading angle, it could move quickly to track the desired path after traveling $1.52\sim3.5m$. For turning section, i.e., the curved path with curvature of 3 m, the vehicle completed its turning securely.

• 전기학회논문지
• /
• 제67권8호
• /
• pp.1089-1098
• /
• 2018

In this paper, we propose an autonomous feeding robot and its obstacle classification system using ultrasonic sensors to secure the driving safety of the robot and efficient feeding operation. The developed feeding robot is verified by operation experiments in a cattle shed. In the proposed classification algorithm, not only the maximum amplitude of the ultrasonic echo signal but also two gradients of the signal and the variation of amplitude are considered as the feature parameters for object classification. The experimental results show the efficiency of the proposed classification method based on the Support Vector Machine, which is able to classify objects or obstacles such as a human, a cow, a fence and a wall.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2009년도 학술대회
• /
• pp.606-610
• /
• 2009

본 논문에서는 단기 서술형 답안의 자동 채점을 위하여 기계 번역 자동 평가에서 널리 사용되는 BLEU(BiLingual Evaluation Understudy)를 활용한 방법을 제안한다. BLEU 는 기계가 번역한 것이 사람이 번역한 것과 비슷할수록 기계번역의 질이 좋을 것이다 라는 것을 가정하여 평가한다. 즉, 특정 문장을 여러 사람이 번역한 문장을 기계가 번역한 문장과 n-gram 방식으로 비교해 점수를 매기는 것이다. 이와 비슷하게 본 연구에서는 여러 개의 정답 문장과 학생의 답안 문장을 BLEU 와 같은 방식으로 상호 비교하여 학생의 답안을 채점하였다. 실험에서는 이러한 채점 방식의 정확도를 평가하기 위하여 사람이 채점한 점수와의 상관관계를 계산하였다.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2004년도 추계학술대회 학술발표 논문집 제14권 제2호
• /
• pp.94-97
• /
• 2004

독립된 자율로봇에서 머신비젼의 구동을 위해 본 논문에서는 DARS(Distributed Autonomous Robotic System)에 적용하기 위한 디지털 이미지 프로세싱을 연구하고, DARS의 개별 로봇에 이를 임베디드화하는 것을 연구한다. 따라서 로봇을 구동하기 위해 필요한 데이터를 CMOS 카메라로부터 수신하여 영상을 스캔한 후, 원영상을 신경망 알고리즘을 통해 클러스터링하여 필요한 데이터를 추출한다. 또 이를 사용자 컴퓨터 단말기 상에 디스플레이하고, 최종적으로 DARS의 자율 이동 로봇이 영상 데이터를 인지하여 특정한 선택 동작을 수행하도록 한다.

• Journal of Biosystems Engineering
• /
• 제21권2호
• /
• pp.167-174
• /
• 1996

A Fuzzy Logic Controller(FLC) was developed for the autonomous operation of speedsprayer in an orchard. The autonomous operation with the FLC was graphically simulated under the real condition of the orchard. Image processing was used to find out the direction of running and four ultrasonic sensors were used to detect obstacles for the running. The simulation results showed that the speedsprayer could be operated autonomously with the FLC combined with the image processing and the ultrasonic sensors.

• 한국정밀공학회지
• /
• 제29권4호
• /
• pp.367-372
• /
• 2012

The biggest factors that lower the machining accuracy are thermal deformation and chatter vibration. In this article, we introduce the study case of technology that can automatically compensate the errors of these factors of a machine during processing on the machine tool's CNC(Computerized Numerical Controller) in real time. This study is related to the detection and compensation of thermal deformation and chatter vibration that can compensate for faster and produce processed goods with more precision by autonomous compensation. In addition, this study is related to the active control of vibration during machining, monitoring of cutting force and auto recognition of machining axes origin. Thus, we attempt to introduce the related contents of the development we have made in this article.

• Journal of Multimedia Information System
• /
• 제7권1호
• /
• pp.41-44
• /
• 2020

In autonomous driving, object recognition based on machine learning is one of the core software technologies. In particular, the object recognition using deep learning becomes an essential element for autonomous driving software to operate. In this paper, we introduce a drivable area detection method based on Region-based CNN model to support autonomous driving. To effectively detect the drivable area, we used the BDD dataset for model training and demonstrated its effectiveness. As a result, our R-CNN model using BDD datasets showed interesting results in training and testing for detection of drivable areas.

• 토지주택연구
• /
• 제11권1호
• /
• pp.103-108
• /
• 2020

Autonomous things system is a technology that judges and acts based on using surrounding information by itself, and it is evaluated as a future technology that can replace the current IoT technology. The current IoT technology is widely used from facility monitoring to machine control but it is shown weakness as a evaluation and prediction technique despite of smart city important technology. In this study, in order to confirm the application of the autonomous things technology, a monitoring system was installed on a real reservoir dam facility and long-term monitoring was performed that the measuring device itself judges and control as a facility monitoring technology. The autonomous things technology was confirmed during 19 months and it is possible to continuous measurement in the same way as current automated instrumentation. In addition, it was confirmed that the ICT device itself could to control autonomously measurement cycle according to the rainfall by itself.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2006년도 춘계학술대회 논문집
• /
• pp.345-346
• /
• 2006

In the future, a machine will be more improved in the form of advanced concept with collaborative ability in M2M(Machine to Machine, Mobile to Machine) environment for u-Manufacturing system. This paper tried to standardize M2M and design advanced concept machine. The M2M is front-end system for implementing autonomous ubiquitous environment. The advanced machine in M2M will be a collaborative machine with knowledge-evolutionary ability such as u-Machine(Ubiquitous machine), Vortal(Vertical Portal) machine and P2P(Peer to Peer) machine. Such advanced concept machines will be the key subject for M2M cooperation.