• Journal of Advanced Research in Ocean Engineering
• /
• v.3 no.3
• /
• pp.136-148
• /
• 2017

The paper presents dual arm ROV manipulation using deep reinforcement learning. The purpose of this underwater manipulator is to investigate and excavate natural resources in ocean, finding lost aircraft blackboxes and for performing other extremely dangerous tasks without endangering humans. This research work emphasizes on a self-learning approach using Deep Reinforcement Learning (DRL). DRL technique allows ROV to learn the policy of performing manipulation task directly, from raw image data. Our proposed architecture maps the visual inputs (images) to control actions (output) and get reward after each action, which allows an agent to learn manipulation skill through trial and error method. We have trained our network in simulation. The raw images and rewards are directly provided by our simple Lua simulator. Our simulator achieve accuracy by considering underwater dynamic environmental conditions. Major goal of this research is to provide a smart self-learning way to achieve manipulation in highly dynamic underwater environment. The results showed that a dual robotic arm trained for a 3DOF movement successfully achieved target reaching task in a 2D space by considering real environmental factor.

• The Bulletin of The Korean Astronomical Society
• /
• v.38 no.2
• /
• pp.83.2-83.2
• /
• 2013

Camera for QUasars in EArly uNiverse (CQUEAN) is an optical CCD camera developed by the Center for the Exploration of the Origin of the Universe (CEOU). In 2010 August, CQUEAN was attached on the 2.1m Otto Struve Telescope at the McDonald Observatory in Texas, USA. As the main purpose of CQUEAN is detecting the Lyman breaks of redshift ~5 quasars, it is sensitive to near-infrared wavelengths (0.7-1.0 ${\mu}m$). For the auto-guiding system, it is using a rotating guide arm to find guide stars on the Cassegrain off-axis focus of the telescope. We plan to upgrade a new filter wheel system consists of a series of narrow band filters. We will install this independent auto-guiding units on the finder scope, which makes rooms on the Cassegrain focal plane of the main telescope. In this presentation we present the system architecture of the CQUEAN Auto-guiding Package (CAP).

• Korean Institute of Interior Design Journal
• /
• v.18 no.2
• /
• pp.3-15
• /
• 2009

In the swirl of change in the world, Media, the development of information technology, and the enterprise-oriented capitalism have tied the world as one unit, and have entered the period of adaptation. With the view of introspection of these concepts, this paper is mainly written by the theory of "matter as the content" with exploration of the design of essence and scientific approach. The design process is the condensational process which is developed by the exchange of reason and sensibility. This process, which is characterized by the objective manipulation of a given condition and the intervention of extrapolation, is the new way of thinking and approach rather than the concept of the way and means. The research of this paper is based on paintings and the analysis of essential experiments in design, the definition of the terms of content in the view of philosophy and design, and the analysis of the cases in architecture and spacial design. As a result, this paper shows that these designs are not just "Simulacre" but the essential eidos, and "content" is the core of these designs which can produce prototype as machine. Also, these designs can relatively be the persuasive methodology for the reflexive modernization.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.36C no.3
• /
• pp.1-7
• /
• 1999

In this research, transformation-based optimization techniques for pipelined designs have been developed. The transformation-based optimization techniques include pipelined architecture transformations and retiming transformations. The new transformation method has the following three features. First, the overall performance of a pipelined system is optimized owing to various transformations including retiming of multiple pipelined blocks. Second, these techniques can be used to search a large solution space by allowing efficient exploration of trade-offs between area and performance. Third, these techniques can be easily extended to a new transformation or algorithm and can be used to optimize memory or bus architectures. Experimental results illustrate that these transformation-based optimization techniques improve area by 21% and performance by 17% on the average for a set of pipelined designs. Especially, the techniques are useful to efficiently explore a large design space.

• Architectural research
• /
• v.13 no.3
• /
• pp.19-29
• /
• 2011

Visibility of a space have been defined in several different ways: such as the axial line covering a convex space, a convex space defining the fattest shape in a space and an Isovist field formed by a field of vision at a given vantage point. Isovist fields are referred to as a descriptive medium to describe a movement by reviewing and analyzing geometric properties in them. Many descriptive methods for analysis of three-dimensional isovist are applied to analyzing the morphological properties in a 3D space more realistically. Although these models are regarded as a more advanced method for describing spatial properties, they have pros and cons such as complex mathematical calculations and somewhat arbitrary calibration in addition to huge consumption of memory space. These difficulties lead to the development of a three-dimensional visual accessibility model that explores the implication of building shape on the calculation of isovist fields drawn on a 2D plane. We propose a conceptual framework of how to measure the isovist field not as a 3D volume but as a combination of 2D plane on the ground with the 3D building shape of it's facade.

• 제어로봇시스템학회:학술대회논문집
• /
• 1999.10a
• /
• pp.9-12
• /
• 1999

This paper presents the development of a control architecture for the autonomous underwater vehicle (AUV) with VARIVEC (variable vector) propeller. Moreover this paper also describes the new technique of controlling the servomotors using the Field Programmable Gate Array (FPGA). The AUVs are being currently used fur various work assignments. For the daily measuring task, conventional AUV are too large and too heavy. A small AUV will be necessary for efficient exploration and investigation of a wide range of a sea. AUVs are in the phase of research and development at present and there are still many problems to be solved such as power resources and underwater data transmission. Further, another important task is to make them smaller and lighter for excellent maneuverability and low power. Our goal is to develop a compact and light AUV having the intelligent capabilities. We employed the VARIVEC propeller system utilizing the radio control helicopter elements, which are swash plate and DC servomotors. The VARIVEC propeller can generate six components including thrust, lateral force and moment by changing periodically the blade angle of the propeller during one revolution. It is possible to reduce the number of propellers, mechanism and hence power sources. Our control tests were carried out in an anechoic tank which suppress the reflecting effects of the wall surface. We tested the developed AUV with required performance. Experimental results indicate the effectiveness of our approach. Control of VARIVEC propeller was realized without any difficulty.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.38 no.4
• /
• pp.277-284
• /
• 2001

Viterbi algorithm is widely used in digital communication system for FEC(forward error correction). Each communication systems based on the Viterbi algorithm use specific Viterbi decoder which has different code parameter values. Even if Viterbi decoder has the same code parameters, it can be varied by the design architecture adopted. We propose the parameterized VHDL model generator for the efficiency of the design. It makes it possible to achieve shorter design time and lower design cost. The model generator searches the design space available and finds out the optimal design point to generate a decoder model.

• Journal of Multimedia Information System
• /
• v.8 no.3
• /
• pp.147-158
• /
• 2021

Versatile Video Coding (VVC) is the latest video coding standard developed by Joint Video Exploration Team (JVET). In VVC, the quadtree plus multi-type tree (QT+MTT) structure of coding unit (CU) partition is adopted, and its computational complexity is considerably high due to the brute-force search for recursive rate-distortion (RD) optimization. In this paper, we aim to reduce the time complexity of inter-picture prediction mode since the inter prediction accounts for a large portion of the total encoding time. The problem can be defined as classifying the split mode of each CU. To classify the split mode effectively, a novel convolutional neural network (CNN) called multi-level tree (MLT-CNN) architecture is introduced. For boosting classification performance, we utilize additional information including inter-picture information while training the CNN. The overall algorithm including the MLT-CNN inference process is implemented on VVC Test Model (VTM) 11.0. The CUs of size 128×128 can be the inputs of the CNN. The sequences are encoded at the random access (RA) configuration with five QP values {22, 27, 32, 37, 42}. The experimental results show that the proposed algorithm can reduce the computational complexity by 11.53% on average, and 26.14% for the maximum with an average 1.01% of the increase in Bjøntegaard delta bit rate (BDBR). Especially, the proposed method shows higher performance on the sequences of the A and B classes, reducing 9.81%~26.14% of encoding time with 0.95%~3.28% of the BDBR increase.

• Journal of Astronomy and Space Sciences
• /
• v.38 no.2
• /
• pp.105-117
• /
• 2021

In this study, a preliminary trajectory design is conducted for a conceptual spacecraft mission to a near-Earth asteroid (NEA) (99942) Apophis, which is expected to pass by Earth merely 32,000 km from the Earth's surface in 2029. This close approach event will provide us with a unique opportunity to study changes induced in asteroids during close approaches to massive bodies, as well as the general properties of NEAs. The conceptual mission is set to arrive at and rendezvous with Apophis in 2028 for an advanced study of the asteroid, and some near-optimal (in terms of fuel consumption) trajectories under this mission architecture are to be investigated using a global optimization algorithm called monotonic basin hopping. It is shown that trajectories with a single swing-by from Venus or Earth, or even simpler ones without gravity assist, are the most feasible. In addition, launch opportunities in 2029 yield another possible strategy of leaving Earth around the 2029 close approach event and simply following the asteroid thereafter, which may be an alternative fuel-efficient option that can be adopted if advanced studies of Apophis are not required.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.24 no.6
• /
• pp.591-601
• /
• 2021

Recently, infrared object detection(IOD) has been extensively studied due to the rapid growth of deep neural networks(DNN). Adversarial attacks using imperceptible perturbation can dramatically deteriorate the performance of DNN. However, most adversarial attack works are focused on visible image recognition(VIR), and there are few methods for IOD. We propose deep learning-based adversarial attacks for IOD by expanding several state-of-the-art adversarial attacks for VIR. We effectively validate our claim through comprehensive experiments on two challenging IOD datasets, including FLIR and MSOD.