• Environmental Engineering Research
• /
• v.22 no.1
• /
• pp.75-86
• /
• 2017

The aerosol content dynamics in a virtual system were investigated. The outcome was extended to monitor the mean concentration diffusion of aerosols in a predefined macro and micro scale. The data set used were wind data set from the automatic weather station; satellite data set from Total Ozone Mapping Spectrometer aerosol index and multi-angle imaging spectroradiometer; ground data set from Aerosol robotic network. The maximum speed of the macro scale (West Africa) was less than 4.4 m/s. This low speed enables the pollutants to acquire maximum range of about 15 km. The heterogeneous nature of aerosols layer in the West African atmosphere creates strange transport pattern caused by multiple refractivity. It is believed that the multiple refractive concepts inhibit aerosol optical depth data retrieval. It was also discovered that the build-up of the purported strange transport pattern with time has enormous potential to influence higher degrees of climatic change in the long term. Even when the African Easterly Jet drives the aerosols layer at about 10 m/s, the interacting layers of aerosols are compelled to mitigate its speed to about 4.2 m/s (macro scale level) and boost its speed to 30 m/s on the micro scale level. Mean concentration diffusion of aerosols was higher in the micro scale than the macro scale level. The minimum aerosol content dynamics for non-decaying, logarithmic decay and exponential decay particulates dispersion is given as 4, 1.4 and 0 respectively.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2006.04a
• /
• pp.395-402
• /
• 2006

In analyzing the nano-scale phenomena or behaviors of nano devices or materials, it is often desirable to deal with more atoms than can be treated only with a full atomistic simulation. However, even now, it is advisable to apply the atomistic simulation to the narrow region where the deformation field changes rapidly but to apply the conventional continuum model to the region far from that region. This equivalent continuum model can be formulated by applying the Cauchy-Born rule to the exact atomistic potential as in the quasicontinuum method. To couple the atomistic model with the equivalent continuum model, continuum displacements are conformed to the molecular displacements at the discrete positions of the atoms within the bridging domain. To satisfy the coupling constraints, we apply the Lagrange multiplier method. The continuum model in the bridging model should be applied on the region where the deformation field changes gradually. Then we can make the nodal spacing in the continuum model be much larger than the atomic spacing. In the first step, we generate the atomic-resolution mesh with the nodal spacing equal to the atomic spacing, and then we eliminate the nodal degrees of freedom adaptively using the node deactivation techniques. We eliminate more DOFs as the regions are more far from the atomistic region. Computing time and computational resources can be greatly reduced by the present node deactivation technique in multi scale analysis.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2005.03a
• /
• pp.461-468
• /
• 2005

In this paper, mixed mode magneto-rheological (MR) damper, which is applicable for vibration control of a small scale multi-story structure, is devised. First, the schematic configurations of the shear, flow, and mixed mode MR dampers are described with design constraints and then the analytical models to predict the field-dependent damping forces are derived for each type. Second, an appropriate size of the mixed mode MR damper is manufactured and its field-dependent damping characteristics are evaluated in time domain. Finally, the performance of the manufactured MR damper which is semi-actively applied to a small scale building excited by earthquake load, is numerically evaluated.

• Journal of Electrical Engineering and Technology
• /
• v.6 no.3
• /
• pp.311-318
• /
• 2011

Small-scale reactors have recently attracted attention as a potential power generation source for the future. The Regional Energy Research Institute for Next Generation is currently developing a small-scale reactor called Regional Energy rX 10 MVA (REX-10). The current paper deals with a power system to be used with small-scale reactors for multi-purpose regional energy systems. This small nuclear system can supply electric and thermal energy like a co-generation system. The electrical model of the REX-10 has been developed as a part of the SCADA system. REX-10's dynamic and electromagnetic performance on the power system is analyzed. Simulations are carried out on a test system based on Ulleung Island's power system to validate REX-10 availability on a power system. RSCAD/RTDS and PSS/E software tools are used for the simulation.

• The Journal of Korea Robotics Society
• /
• v.12 no.3
• /
• pp.313-321
• /
• 2017

One of the most frequently performed tasks in human-robot interaction (HRI), intelligent vehicles, and security systems is face related applications such as face recognition, facial expression recognition, driver state monitoring, and gaze estimation. In these applications, accurate head pose estimation is an important issue. However, conventional methods have been lacking in accuracy, robustness or processing speed in practical use. In this paper, we propose a novel method for estimating head pose with a monocular camera. The proposed algorithm is based on a deep neural network for multi-task learning using a small grayscale image. This network jointly detects multi-view faces and estimates head pose in hard environmental conditions such as illumination change and large pose change. The proposed framework quantitatively and qualitatively outperforms the state-of-the-art method with an average head pose mean error of less than $4.5^{\circ}$ in real-time.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10a
• /
• pp.95-98
• /
• 1996

Universal Learning Network(U.L.N.), which can model and control the large scale complicated systems naturally, consists of nonlinearly operated nodes and multi-branches that may have arbitrary time delays including zero or minus ones. Therefore, U.L.N. can be applied to many kinds of systems which are difficult to be expressed by ordinary first order difference equations with one sampling time delay. It has been already reported that learning algorithm of parameter variables in U.L.N. by forward and backward propagation is useful for modeling, managing and controlling of the large scale complicated systems such as industrial plants, economic, social and life phenomena. But, in the previous learning algorithm of U.L.N., time delays between the nodes were fixed, in other words, criterion function of U.L.N. was improved by adjusting only parameter variables. In this paper, a new learning algorithm is proposed, where not only parameter variables but also time delays between the nodes can be adjusted. Because time delays are integral numbers, adjustment of time delays can be carried out by a kind of random search procedure which executes intensified and diversified search in a single framework.

• Journal of the Korea Convergence Society
• /
• v.12 no.11
• /
• pp.45-51
• /
• 2021

With the development of deep learning, semantic segmentation methods are being studied in various fields. There is a problem that segmenation accuracy drops in fields that require accuracy such as medical image analysis. In this paper, we improved PSPNet, which is a deep learning based segmentation method to minimized the loss of features during semantic segmentation. Conventional deep learning based segmentation methods result in lower resolution and loss of object features during feature extraction and compression. Due to these losses, the edge and the internal information of the object are lost, and there is a problem that the accuracy at the time of object segmentation is lowered. To solve these problems, we improved PSPNet, which is a semantic segmentation model. The multi-scale attention proposed to the conventional PSPNet was added to prevent feature loss of objects. The feature purification process was performed by applying the attention method to the conventional PPM module. By suppressing unnecessary feature information, eadg and texture information was improved. The proposed method trained on the Cityscapes dataset and use the segmentation index MIoU for quantitative evaluation. As a result of the experiment, the segmentation accuracy was improved by about 1.5% compared to the conventional PSPNet.

• Journal of Information Display
• /
• v.2 no.2
• /
• pp.13-21
• /
• 2001

We propose a universal image processor for a-Si TFT LCD of UXGA class that can display the full screen on the LCD panel with low resolution of video sources such as NTSC, VGA, SVGA, XGA, and SXGA by using the proposed interpolation filter. In addition, we propose a real-time contrast controller for image improvement of multi-gray scale image. The operation of the proposed methods has been verified using Synopsys VHDL and computer simulation. Results show that the proposed methods might be suitable for a UXGA LCD controller for real-time image improvement.

• Journal of the Korean housing association
• /
• v.25 no.1
• /
• pp.15-22
• /
• 2014

In the domestic housing market, uncompleted lot-solid apartments and unoccupied residences are increasing in the medium-large scale area whose rents and transactions have been unstable, as going through the financial crisis and economical recession. Demands for small-scale housings are steadily increasing by practical consumer group who don't need large places, which are often not needed for the current people who are spending most of their time outside of their houses. Therefore, this study proposes the useful methods for spatial planning for small-scale houses for multiple households, including variable form, levels form, multi-level form, and smart-sizing. For efficient spatial planning for that purpose, here, small-scale housings are not limited to places for families with 1-2 members, but considered as one of broader approaches to graft various households and families. In addition, we examined the existed trends of researches about spatial utility of small-scale places, and did the research about the status of small-scale housing developments, and then proposed the direction for future small-scale housings by analyzing the advanced examples.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.33-42
• /
• 2022

In this paper, we proposed a deep learning model based on multi-scale dilated convolution feature fusion for the segmentation of black ice region in road image to send black ice warning to drivers in real time. In the proposed multi-scale dilated convolution feature fusion network, different dilated ratio convolutions are connected in parallel in the encoder blocks, and different dilated ratios are used in different resolution feature maps, and multi-layer feature information are fused together. The multi-scale dilated convolution feature fusion improves the performance by diversifying and expending the receptive field of the network and by preserving detailed space information and enhancing the effectiveness of diated convolutions. The performance of the proposed network model was gradually improved with the increase of the number of dilated convolution branch. The mIoU value of the proposed method is 96.46%, which was higher than the existing networks such as U-Net, FCN, PSPNet, ENet, LinkNet. The parameter was 1,858K, which was 6 times smaller than the existing LinkNet model. From the experimental results of Jetson Nano, the FPS of the proposed method was 3.63, which can realize segmentation of black ice field in real time.