• 한국위성정보통신학회논문지
• /
• 제10권2호
• /
• pp.19-23
• /
• 2015

위성통신서비스의 가용율 및 시스템 throughput 향상을 위해 사용하는 리턴링크 ACM(Adaptive Coding & Modulation)의 원리를 기술하였고, LMS(Least Mean Square) 기반 적응형 필터를 이용한 채널 예측 및 단말의 전송 MODCOD(Modulation & Code rate) 결정 알고리즘의 최적화 과정을 서술하였다. 시뮬레이션 결과 LMS 알고리즘은 필터 계수가 2차이고, ${\mu}$(step size) 값이 0.00026인 경우 MMSE(Minimum Mean Square Error)가 최소임을 알 수 있다. 이때 MODCOD 결정 알고리즘을 위한 SNR 마진이 0.3dB일 경우 MODCOD 결정 오차를 최소화 할 수 있음을 확인하였다.

• International Journal of Advanced Culture Technology
• /
• 제5권2호
• /
• pp.74-81
• /
• 2017

Recently, there are many companies that use systems based on artificial intelligence. The accuracy of artificial intelligence depends on the amount of learning data and the appropriate algorithm. However, it is not easy to obtain learning data with a large number of entity. Less data set have large generalization errors due to overfitting. In order to minimize this generalization error, this study proposed DGA(Dropout Genetic Algorithm) which can expect relatively high accuracy even though data with a less data set is applied to machine learning based genetic algorithm to deep learning based dropout. The idea of this paper is to determine the active state of the nodes. Using Gradient about loss function, A new fitness function is defined. Proposed Algorithm DGA is supplementing stochastic inconsistency about Dropout. Also DGA solved problem by the complexity of the fitness function and expression range of the model about Genetic Algorithm As a result of experiments using MNIST data proposed algorithm accuracy is 75.3%. Using only Dropout algorithm accuracy is 41.4%. It is shown that DGA is better than using only dropout.

• Current Optics and Photonics
• /
• 제3권6호
• /
• pp.531-540
• /
• 2019

Spectral Amplitude Coding Optical Code Division Multiple Access (SAC OCDMA) is an advanced technique in asynchronous environments. This paper proposes design and implementation of a novel quantum logic gate (QLG) code, with code construction algorithm generated without following any code mapping procedures for SAC system. The proposed code has a unitary matrices property with maximum overlap of one chip for various clients and no overlaps in spectra for the rest of the subscribers. Results indicate that a single algorithm produces the same length increment for codes with weight greater than two and follows the same signal to noise ratio (SNR) and bit error rate (BER) calculations for a higher number of users. This paper further examines the performance of a QLG code based SAC-OCDMA system with NAND and direct detection techniques. BER analysis was carried out for the proposed code and results were compared with existing MDW, RD and GMP codes. We demonstrate that the QLG code based system performs better in terms of cardinality, which is followed by improved BER. Numerical analysis reveals that for error free transmission (10^-9), the suggested code supports approximately 170 users with code weight 4. Our results also conclude that the proposed code provides improvement in the code construction, cross-correlation and minimization of noises.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권6호
• /
• pp.2302-2316
• /
• 2015

We propose an efficient method of extracting targets within a region of interest in non-homogeneous infrared images by using a principal component analysis (PCA) plane and adaptive Gaussian kernel. Existing approaches for extracting targets have been limited to using only the intensity values of the pixels in a target region. However, it is difficult to extract the target regions effectively because the intensity values of the target region are mixed with the background intensity values. To overcome this problem, we propose a novel PCA based approach consisting of three steps. In the first step, we apply a PCA technique minimizing the total least-square errors of an IR image. In the second step, we generate a binary image that consists of pixels with higher values than the plane, and then calculate the second derivative of the sum of the square errors (SDSSE). In the final step, an iteration is performed until the convergence criteria is met, including the SDSSE, angle and labeling value. Therefore, a Gaussian kernel is weighted in addition to the PCA plane with the non-removed data from the previous step. Experimental results show that the proposed method achieves better segmentation performance than the existing method.

• 한국전자파학회논문지
• /
• 제29권4호
• /
• pp.316-319
• /
• 2018

전방 이동지상 표적 탐지를 위해 미사일은 FMCW 레이다를 통해 획득한 SAR 영상을 활용할 수 있다. 하지만 미사일 이동 과정에서 난기류 또는 비행경로 상태에 따른 요동 오차로 인해 SAR 영상의 품질이 떨어지게 된다. 본 논문에서는 이러한 전방을 바라보는 SAR 영상의 요동 오차를 보상하기 위한 엔트로피 최소 자동초점기법을 제안한다. 특히, 전방을 바라보는 레이다 특성 상, 요동 오차는 2차원의 형태(방위각 및 시간 축)로 SAR 영상에 영향을 미치며, 이를 보상하기 위해 2차원 자동초점기법을 제시한다.

• 한국통신학회논문지
• /
• 제36권4A호
• /
• pp.337-344
• /
• 2011

본 논문은 모든 노드가 다중 안테나를 갖는 다중 안테나 (MIMO, multiple-input and multiple-output) 공간 다중화 (SM, spatial multiplexing) 릴레이 시스템을 비트 오율 (BER, bit error rate) 관점에서 연구한다. 제한된 전력 자원을 효율적으로 이용하기 위해서는 각 노드와 안테나에서 최적화된 전력 할당 전략이 필요하다. 본 논문은 이런 관점에서 다중 안테나 공간 다중화 릴레이 시스템을 위한 비트 오율 최소화에 기반을 둔 전력 할당 알고리즘을 제안한다. 제안된 알고리즘은 평균 비트 오율을 직접 최소화하여 얻어지며, 노드 간 (inter-node) 전력 할당 알고리즘과 안테나 간 (inter-antenna) 전력 할당 알고리즘으로 구성된다. 비트 오율 성능에 있어서, 기존의 균등 전력 할당 (EPA, equal power allocation) 알고리즘보다 추가적인 전력 소비 없이도 월등한 성능을 보인다.

• 대한기계학회논문집
• /
• 제11권3호
• /
• pp.509-518
• /
• 1987

본 연구에서는 냉간 축대칭 전방 압출에 가중잔류항법을 적용하여 재료의 가 공 경화 및 강소성 경계를 고려하는 프로그램을 개발하여 변형도, 응력, 변형력, 강소 성 경계등을 FEM과 동일한 조건에서 비교 해석하고 다른 공정에 적용할 수 있게 하고 또한 곡면다이와 원추형다이를 설계 제작하여 다이의 형상과 단면 감소율이 변형도와 응력 분포에 미치는 영향을 검토하고 압출된 제품의 성질을 분석하여 실제 공정에 이 바지하며 이론 계산과 실험을 비교함이 목적이다.

• 한국소음진동공학회논문집
• /
• 제23권11호
• /
• pp.943-950
• /
• 2013

This paper discusses a harmonic response estimation method on the L$\acute{e}$vy plate with two opposite edges simply supported and the other two edges having free boundary conditions. Since the equation of motion of the plate is not self-adjoint, the modes are not orthogonal to each other on the domain. Noting that the L$\acute{e}$vy plate can be expressed using one term sinusoidal function that is orthogonal to other sinusoidal functions, this paper suggested the calculation method that is equivalent to finding a least square error minimization solution of the finite number of algebraic equations. Example problems subjected to a distributed area loading and a distributed line loading are defined and their solutions are provided. The solutions are compared to those of the commercial code, ANSYS. According to the verification results, it is expected that the suggested method will be useful to predict the forced response on the L$\acute{e}$vy plate with the distributed area or line loading conditions.

• 한국시뮬레이션학회논문지
• /
• 제26권3호
• /
• pp.125-139
• /
• 2017

최근 응급의료 서비스의 질 제고에 관심이 높아지고 있는 가운데 응급의료 프로세스의 혁신에 많은 노력을 기울이고 있다. ICT기술의 급속적인 진전에 의해 응급의료 프로세스의 자동화 또는 지능화가 가속화되고 있다. 본 연구는 자원 활용 최적화, 인적오류 최소화 그리고 진료 예측가능성 제고를 고려한 실시간데이터 기반 응급실 운영 방안을 제안한다. 응급실 운영지수-응급 케어지수, 체류 단축지수, 인적오류 유발지수, 대기 인내지수-를 개발하고, 이를 기반으로 한 응급실 운영규칙을 제시한다. 가상의 축소 응급실을 대상으로 시뮬레이션을 실시하여 제안한 운영규칙의 효과성을 검증하였다. 시뮬레이션 결과 응급실 체류시간에서 우수한 성능을 보였다.

• International journal of advanced smart convergence
• /
• 제8권1호
• /
• pp.141-146
• /
• 2019

These days, there are many applications using neural networks as parts of their system. On the other hand, adversarial examples have become an important issue concerining the security of neural networks. A classifier in neural networks can be fooled and make it miss-classified by adversarial examples. There are many research to encounter adversarial examples by using denoising methods. Some of them using GAN (Generative Adversarial Network) in order to remove adversarial noise from input images. By producing an image from generator network that is close enough to the original clean image, the adversarial examples effects can be reduced. However, there is a chance when adversarial noise can survive the approximation process because it is not like a normal noise. In this chance, we propose a research that utilizes high-level representation in the classifier by combining GAN network with a trained U-Net network. This approach focuses on minimizing the loss function on high representation terms, in order to minimize the difference between the high representation level of the clean data and the approximated output of the noisy data in the training dataset. Furthermore, the generated output is checked whether it shows minimum error compared to true label or not. U-Net network is trained with true label to make sure the generated output gives minimum error in the end. At last, the remaining adversarial noise that still exist after low-level approximation can be removed with the U-Net, because of the minimization on high representation terms.