http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Van-Sang Doan,김동성 한국통신학회 2020 ICT Express Vol.6 No.2
Up to date, investigation of direction of arrival (DOA) gains more effort on multiple signals rather than coherent signals, particularly coherent signals measured by non-uniform and circular arrays — which still remains room for exploration. In this study, a new approach using element transposition of the covariance matrix combined with the MUSIC algorithm was proposed to estimate DOAs of multiple coherent and non-coherent signals by an arbitrary symmetric antenna array. A balanced property of phase differences of the symmetric antenna array was discovered to de-correlate the received signals, and simulation was performed for linear and circular arrays. Here, we established a new method to estimate and distinguish the DOAs of both non-coherent and coherent signals with higher accuracy and resolution than other conventional methods.
Modified ShuffleNet-based Radar Signal Classification of Electronic Intelligence System
Van Sang Doan,Dong-Seong Kim(김동성) 한국통신학회 2020 한국통신학회 학술대회논문집 Vol.2020 No.2
ShuffleNet, an extremely efficient convolutional neural network for mobile devices due to its lightweight structure, is modified and improved in this work to apply for classifying the radar signal waveforms with signal surveillance and reconnaissance purposes of passive radar systems. Unlike traditional hand-craft extraction-based methods, which use measured parameters (key features) of signal such as carrier frequency, signal power, pulse repetition frequency, pulse width, etc., for classification, the deep neural network can directly operate with the raw signal in the time domain to automatically classify them at the output. In this study, simulation is performed to validate the efficiency of the proposed neural network model, which can be also taken into comparison with state-of-the-art other models in our future work.
Van Thuan, Doan,Nguyen, Tri Khoa,Kim, Soon-Wook,Chung, Jin Suk,Hur, Seung Hyun,Kim, Eui Jung,Hahn, Sung Hong,Wang, Mingsong Elsevier 2017 CATALYSIS COMMUNICATIONS - Vol.101 No.-
<P><B>Abstract</B></P> <P>Oval-shaped graphene/ZnO quantum hybrid (GZQH) is synthesized via chemical-hydrothermal method and tested for the photoenhanced selective reduction of nitroarenes. A facile molecular fusion process is employed to produce graphene quantum dots (GQDs) from pyrene, which is followed by hydrothermal treatment with embryonic ZnO quantum dots (5nm in size) to yield the GZQHs. Zn<SUP>2+</SUP> ions on ZnO embryo react with a functional group of graphene, which forms Zn-O-C bonding leading to highly crystalline quantum hybrids with uniform interface. The GZQHs have a quenched photoluminescence intensity as compared to the GQDs (2nm in size) due to electron transfer at the graphene-ZnO interface. Hydrogen molecules dissociate into hydrogen atoms by photogenerated electrons which transfer and perturb at the interface under UV irradiation. The GZQHs exhibit an excellent UV-induced catalytic performance for the selective reduction of nitroarenes. The effect of ZnO:graphene ratio on reduction reaction rate constant is also investigated.</P> <P><B>Highlight</B></P> <P> <UL> <LI> Excellent hydrogenation reduction of noble-metal-absent catalyst: Graphene and ZnO. </LI> <LI> Graphene and zinc oxide composite at the quantum scale particle synthesis. </LI> <LI> Hydrogen dissociation at G/ZnO interface and spillover to graphene. </LI> <LI> UV stimulated electron transfer and perturbation within Gr/ZnO. </LI> <LI> Catalytic dependence of NaBH<SUB>4</SUB> and Gr/ZnO ratio and reduction of different nitroarenes. </LI> </UL> </P> <P><B>Graphical abstract</B></P> <P>[DISPLAY OMISSION]</P>
Improving Deep CNN-Based Radar Target Classification Performance by Applying a Denoise Filter
Nguyen Van-Tra,Vu Chi-Thanh,Doan Van Sang 한국전자파학회 2024 Journal of Electromagnetic Engineering and Science Vol.24 No.2
This paper presents a novel method for removing noise from range-Doppler images by using a filter prior to conducting target classification using a deep neural network. Specifically, Kuan, Frost, and Lee filters are employed to eliminate speckle noise components from radar data images. Furthermore, a neural network that combines residual and inception blocks (RINet) is proposed. The RINet model is trained and tested on the RAD-DAR dataset—a collection of range-Doppler feature maps. The analysis results show that the application of a Lee filter with a window size of 7 in the RAD-DAR dataset demonstrates the most improvement in the model’s classification performance. On applying this noise filter to the dataset, the RINet model successfully classified radar targets, exhibiting a 4.51% increase in accuracy and a 14.07% decrease in loss compared to the classification results achieved for the original data. Furthermore, a comparison of the RINet model with the noise filtering solution with five other networks was conducted, the results of which show that the proposed model significantly outperforms the others.
Low-Complexity Convolution Neural Network for Accurate Arrival-of-Angle Estimation in Low SNR
Rubina Akter,Van-Sang Doan,Jae-Min Lee,Dong-Seong Kim 한국통신학회 2022 한국통신학회 학술대회논문집 Vol.2022 No.2
This paper presents a low-complexity and robustly applied convolution neural network for angle of arrival (AOA) estimation for object localization. The deep network named LRCANet is cleverly designed using asymmetric convolution kernels specified in multiple residual blocks to successively learn the multi-scale feature maps of a time-variant signal correlation and the spatial relations of different antennas. According to the empirical results, LRCANet outperforms other existing models with a classification rate of 96.35% and an angle error of 0.2<SUP>0</SUP> RMSE at +5 dB signal-to-noise ratio.
Enhanced photostability in polymer solar cells achieved with modified electron transport layer
Rasool, Shafket,Van Doan, Vu,Lee, Hang Ken,Lee, Sang Kyu,Lee, Jong-Cheol,Moon, Sang-Jin,So, Won Wook,Song, Chang Eun,Shin, Won Suk Elsevier 2019 THIN SOLID FILMS - Vol.669 No.-
<P><B>Abstract</B></P> <P>Photostability of the polymer solar cells (PSCs) remains a challenge to attain, as number of factors including the electron transport layers contribute to the degradation of PSCs when they are tested under 1 sun illumination, along with heat and humidity. Especially electron transport layers (ETLs) have considerable contribution to the overall degradations in these solar cells. Most studied ETL into the fabrication of inverted PSCs is zinc oxide (ZnO) yet the photostability of these PSCs is limited. This degradation most probably occurred due to the direct contact of zinc metal with the photoactive layer. This interface between ZnO and photoactive layer results in the initiation of degradations in PSCs. Keeping in view of this issue, we have modified the ZnO ETL by mixing ZnO nano particles with ethoxylated polyethyleneimine (PEIE) and then passivating this modified ZnO (m-ZnO) with ultrathin PEIE buffer layer. The power conversion efficiency is not affected by this approach, but when PSCs were subjected to 1 sun illumination, these m-ZnO/PEIE based PSCs have shown the enhanced light soaking (LS) stability. These findings indicate that optimizing the interface between the photoactive layer and the electron transport layer can lead to enhanced LS stability.</P> <P><B>Highlights</B></P> <P> <UL> <LI> Ethoxylated polyethylenimine (PEIE) passivated ZnO (m-ZnO) as electron transport layer (ETL). </LI> <LI> The direct contact of ZnO with photoactive layer is avoided. </LI> <LI> The m-ZnO/PEIE layer help in enhanced light soaking (LS) stability. </LI> <LI> The m-ZnO/PEIE layer has nitrogen enriched surface. </LI> <LI> The nitrogen enriched ETL surface contributes to enhanced LS stability. </LI> </UL> </P>
Le-Hai Cao,Hoang Van-Phuc,Doan Van Sang,Le Dai Phong 한국전자파학회 2022 Journal of Electromagnetic Engineering and Science Vol.22 No.3
Hand gesture recognition is an efficient and practical solution for the non-contact human–machine interaction in smart devices. To date, vision-based methods are widely used in this research area, but they are susceptible to light conditions. To address this issue, radar-based gesture recognition using micro-Doppler signatures can be applied as an alternative. Accordingly, the use of a novel densely convolutional neural network model, Dop-DenseNet, is proposed in this paper for improving hand gesture recognition in terms of classification accuracy and latency. The model was designed with cross or skip connections in a dense architecture so that the former features, which can be lost in the forward-propagation process, can be reused. We evaluated our model with different numbers of filter channels and experimented with it using the Dop-Net dataset, with different time lengths of input data. As a result, it was found that the model with 64 3 × 3 filters and 200 time bins of micro-Doppler spectrogram data could achieve the best performance trade-off, with 99.87% classification accuracy and 3.1 ms latency. In comparison, our model remarkably outperformed the selected state-of-the-art neural networks (GoogLeNet, Res- Net-50, NasNet-Mobile, and MobileNet-V2) using the same Dop-Net dataset.
Hoang, Quoc Viet,Rasool, Shafket,Oh, Sora,Van Vu, Doan,Kim, Da Hun,Lee, Hang Ken,Song, Chang Eun,Lee, Sang Kyu,Lee, Jong-Cheol,Moon, Sang-Jin,Shin, Won Suk Royal Society of Chemistry 2017 Journal of Materials Chemistry C Vol.5 No.31
<▼1><P>A newly synthesized small molecule donor, BDTT-2DPPBFu, is designed for incorporation into the photo-active layer in solution-processed small molecule organic photovoltaics (SM-OPVs).</P></▼1><▼2><P>A newly synthesized small molecule donor, BDTT-2DPPBFu, is designed for incorporation into the photo-active layer in solution-processed small molecule organic photovoltaics (SM-OPVs). The effects of a solvent additive (1-chloronaphthalene, CN) on the bulk heterojunction morphology of a BDTT-2DPPBFu donor:fullerene derivative acceptor (PC71BM) are investigated and correlated with the device performance. A TEM analysis revealed that the nanomorphology of the SM-OPVs evolved dramatically with increasing solvent additive volume. With an optimum concentration of 2.5 vol% CN, a nanomorphology with a fibrillar and interpenetrating network on the order of exciton diffusion length (∼10–15 nm) is formed, resulting in improvement of the short-circuit current density, fill factor, and power conversion efficiency. Further GIWAXS studies have shown that the π–π stacking distance of BDTT-2DPPBFu small molecules is reduced and the inter-mixing of PC71BM within the BDTT-2DPPBFu-rich phase in the photo-active layer processed from 2.5 vol% CN is increased, leading to enhanced charge transport and reduced charge recombination and transfer resistance. The present work provides important progress in the nanomorphology of the photo-active layer, which can be effectively tuned <I>via</I> the use of a solvent additive in SM-OPVs.</P></▼2>