• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.798-802
• /
• 2003

The robustness and reliability of vision algorithms is the key issue in robotic research and industrial applications. In this paper, the robust real time visual tracking in complex scene is considered. A common approach to increase robustness of a tracking system is to use different models (CAD model etc.) known a priori. Also fusion of multiple features facilitates robust detection and tracking of objects in scenes of realistic complexity. Because voting is a very simple or no model is needed for fusion, voting-based fusion of cues is applied. The approach for this algorithm is tested in a 3D Cartesian robot which tracks a toy vehicle moving along 3D rail, and the Kalman filter is used to estimate the motion parameters, namely the system state vector of moving object with unknown dynamics. Experimental results show that fusion of cues and motion estimation in a tracking system has a robust performance.

• Proceedings of the IEEK Conference
• /
• 2003.07c
• /
• pp.2717-2720
• /
• 2003

A mathematical framework for deriving acceleration bounds from given joint torque limits of two cooperating robots are described in this paper. Especially when the torque limits are given in 2-norm, the resultant geometrical configuration is ellipsoid(the ellipsoid is often called manipulability ellipsoid in many works). At first, the mathematical derivation starts from the dynamics of both object and robots as well as the kinematics of the robots, and is finally arranged in a form of equation relating joint torques to object acceleration through a complete constraint contact(or “very-soft contact”). To show the usefulness of the proposed method, two examples are included, and especially the case where friction effects the ellipsoid shape is also considered In the example.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.2
• /
• pp.693-709
• /
• 2018

In this paper, we propose a compact Convolutional Neural Network (CNN)-based tracker in conjunction with a particle filter architecture, in which the CNN model operates as an accurate candidates estimator, while the particle filter predicts the target motion dynamics, lowering the overall number of calculations and refines the resulting target bounding box. Experiments were conducted on the Online Object Tracking Benchmark (OTB) [34] dataset and comparison analysis in respect to other state-of-art has been performed based on accuracy and precision, indicating that the proposed algorithm outperforms all state-of-the-art trackers included in the OTB dataset, specifically, TLD [16], MIL [1], SCM [36] and ASLA [15]. Also, a comprehensive speed performance analysis showed average frames per second (FPS) among the top-10 trackers from the OTB dataset [34].

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.443-446
• /
• 2002

Wake downstream of an object in the stratified flow has been of long-standing interest in fluid dynamics because of its similarity to geophysical flow over topographical terrains and more recently, concerns about the wake left behind a body moving through the ocean thermocline. Decades of studies of geophysical flow have unveiled that the flow downstream of obstacles in stratified flow consists of attached wake and strong internal waves, or separated, fluctuating wake and persistent late wakes, all of which depend on the flow conditions. Among unique and interesting characteristics of the stratified flow past obstacles is the generation of coherent vortex structure in the late wake far downstream of the object. Without the density stratification, the flow field downstream becomes undisturbed after relatively fast diminishing of the near wake. However, no matter how small the stratification is, the flow field downstream self-develops coherent vortex structures even after diminishing of the near wake. This paper present a computational approach to simulate the generation mechanism of the coherent vortex and analysis of the vortical structure.

• Journal of the Korean Society of Industry Convergence
• /
• v.18 no.3
• /
• pp.167-172
• /
• 2015

In this paper, we propose a new approach to design and control a smart gripper of robot system. A control method for flexible grasping a object in partially unknown environment was proposed, where a proximate sensor detecting the distance between the fingertip and object was used. Based on the proximate sensor signal the finger motion controller could plan the grasping process divided in three phases. The first step is scanning process which two first joints were moved to mid-position of the detected range by a state-variable feedback position controller, after the scanning was finished. The contact force of fingertip was then controlled using the detection sensor of the servo controller for finger joint control. The proposed grasping planning was tested on rectangular bar.

• 한국전산유체공학회:학술대회논문집
• /
• 2011.05a
• /
• pp.483-487
• /
• 2011

In the present study, a CFD program based on a finite volume method was developed by using an unstructured polyhedral grid system for the accurate simulation with the complex geometry of computational domain. To simulate the transient flow induced by the moving solid object, the program used a fractional step method and a ALE (Algebric Lagrangian-Eulerian) method. The grid deformation for the moving of solid object were performed with a spring analogy based on the center coordinate of each computational grid. To verify the present program with these methodologies, the numerical results of the flow around the fixed and oscillating circular cylinder were compared with the previous numerical results.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.474-474
• /
• 2000

This paper shows that a system with two-link arm can obtain hand reaching movement to a target object projected on a visual sensor by reinforcement learning using a layered neural network. The reinforcement signal, which is an only signal from the environment, is given to the system only when the hand reaches the target object. The neural network computes two joint torques from visual sensory signals, joint angles, and joint angular velocities considering the urn dynamics. It is known that the trajectory of the voluntary movement o( human hand reaching is almost straight, and the hand velocity changes like bell-shape. Although there are some exceptions, the properties of the trajectories obtained by the reinforcement learning are somewhat similar to the experimental result of the human hand reaching movement.

• Proceeding of EDISON Challenge
• /
• 2014.03a
• /
• pp.550-555
• /
• 2014

전익기는 전쟁 중 등장해 활약했다는 점과 그 독특한 형태로 유명세를 타기 시작한 비행체의 한 종류이다. 공기역학적으로 동일한 속도로 비행할 때 일반적인 항공기와 비교해서 적은 동력으로 더 멀리 날수 있는 전익기의 특성과 그 발전방향을 보았을 때 동체의 익면적이 넓어 생기는 특성을 극대화 시키면서 더욱 독특한 형태의 비행체로 나타나는 모습이 원반형 비행체이다. 본 연구는 EDISON Simulation을 활용한 두 비행체의 공력특성 비교를 통해 실제로 원반형 비행체가 많이 쓰이지 않는 이유와 그 장단점에 대한 데이터를 확보했으며 특히 원반형 비행체의 경우 Cockpit 유무와 그 크기에 따라 실속각이 커지는 것을 확인하였다.

• Proceedings of the KSME Conference
• /
• 2001.06b
• /
• pp.105-109
• /
• 2001

This paper discuss the real-time self-tuning learning control based on evolutionary computation, which proves its the superiority in the finding of the optimal solution at the off-line learning method. The individuals are reduced in order to learn the evolutionary strategy in real-time, and new method that guarantee the convergence of evolutionary mutations are proposed. It possible to control the control object varied as time changes. As the state value of the control object is generated, applied evolutionary strategy each sampling time because the learning process of an estimation, selection, mutation in real-time. These algorithms can be applied, the people who do not have knowledge about the technical tuning of dynamic systems could design the controller or problems in which the characteristics of the system dynamics are slightly varied as time changes.

• Proceedings of the KSME Conference
• /
• 2001.06b
• /
• pp.724-729
• /
• 2001

This paper discusses the composition of the theory of reinforcement learning, which is applied in real-time learning, and evolutionary strategy, which proves its the superiority in the finding of the optimal solution at the off-line learning method. The individuals are reduced in order to learn the evolutionary strategy in real-time, and new method that guarantee the convergence of evolutionary mutations are proposed. It possible to control the control object varied as time changes. As the state value of the control object is generated, applied evolutionary strategy each sampling time because the learning process of an estimation, selection, mutation in real-time. These algorithms can be applied, the people who do not have knowledge about the technical tuning of dynamic systems could design the controller or problems in which the characteristics of the system dynamics are slightly varied as time changes. In the future, studies are needed on the proof of the theory through experiments and the characteristic considerations of the robustness against the outside disturbances.