This paper describes an application of artificial neural network to diagnose the defects of rotating machiner. Induction motor was used to the object of defect diagnosis. For defect diagnosis, the frequency spectrum of vibration was utilized. Learning method of applied neural network was back propagation. Neural network has following advantage; Once it has been learned, inference time is very short and it can provide a reasonable conclusion regardless of insufficient input data. So, this defect diagnosis system can be used superiorly to rule based expert system as quality inspection of rotating machinery in the shop.

With the introduction of the tele-monitoring system (TMS) in South Korea, monitoring of the concentration of pollutants discharged from nationwide water quality TMS attachments is possible. In addition, the Ministry of Environment is implementing a smart sewage system program that combines ICT technology with wastewater treatment plants. Thus, many institutions are adopting the automatic operation technique which uses process operation factors and TMS data of sewage treatment plants. As a part of the preliminary study, a multilayer perceptron (MLP) analysis method was applied to TMS data to identify predictability degree. TMS data were designated as independent variables, and each pollutant was considered as an independent variables. To verify the validity of the prediction, root mean square error analysis was conducted. TMS data from two public sewage treatment plants in Chungnam were used. The values of RMSE in SS, T-N, and COD predictions (excluding T-P) in treatment plant A showed an error range of 10%, and in the case of treatment plant B, all items showed an error exceeding 20%. If the total amount of data used MLP analysis increases, the predictability of MLP analysis is expected to increase further.

The aim of this study is to develop basic artificial neural network models in forecasting the in-sample gross domestic product (GDP) of Malaysia. GDP is one of the main indicators in presenting the macro economic condition of a country as set by the world authority bodies such as the World Bank. Hence, this study uses an artificial neural network-based approach to make predictions concerning the economic growth of Malaysia. This method has been proposed due to its ability to overcome multicollinearity among variables, as well as the ability to cope with non-linear problems in Malaysia’s growth data. The selected inputs and outputs are based on the previous literatures as well as the economic growth theory. Therefore, the selected inputs are exports, imports, private consumption, government expenditure, consumer price index (CPI), inflation rate, foreign direct investment (FDI) and money supply, which includes M1 and M2. Whilst, the output is real gross domestic product growth rate. The results of this study showed that the neural network method gives the smallest value of mean error which is 0.81 percent with a total difference of 0.70 percent. This implies that the neural network model is appropriate and is a relevant method in forecasting the economic growth of Malaysia.

항만의 주요 정책 및 향후 운영계획 수립 시 정확한 물동량 예측에 관한 연구는 매우 중요하며 이러한 중요성으로 인해 관련 연구가 활발히 수행되고 있다. 본 논문에서는 국내 최대 석탄 및 철광석 처리 항만인 광양항을 대상으로 단계적 회귀분석과 인공신경망모형을 활용하여 모형간 예측력을 비교하였다. 2009년 1월부터 2019년 1월까지 총 121개월의 월별자료를 활용하였으며 석탄 및 철광석 물동량에 영향을 주는 요인을 선정하여 공급관련요인과 시장·경제관련요인으로 분류하였다. 단계적 회귀분석 결과, 광양항 석탄 물동량 예측모형의 경우, 입항선박 톤수, 석탄가격 및 대미환율이 최종변수로 선정되었고 철광석 물동량 예측모형의 경우, 입항선박 톤수, 철광석가격이 최종변수로 선정되었다. 인공신경망모형의 경우, 모델 성능에 영향을 미치는 다양한 Hyper-parameters를 조정하며 최적 모델을 선정하는 시행착오법을 사용하였다. 분석결과 인공신경망모형이 단계적 회귀분석에 비해 우수한 예측성능을 나타내었으며 예측 모형별 예측값과 실측값을 그래프 상 비교 시에도 인공신경망모형이 단계적 회귀분석에 비해 고·저점을 유사하게 나타냈다.

Global warming due to the increase of greenhouse gases may significantly affect various aspects of the Earth’s environment and human life. In particular, the impacts of climate change on agriculture would be severe, leading to damages to crop yields. This paper examines the experimental prediction of rice yield in China using DNN (deep neural network) and climate model data for the period between 1979 and 2009. The DNN model built through the process of hyperparameter optimization can mitigate an overfitting problem and cope with outlier cases. Our model showed approximately 38.7% improved accuracy than the MLR (multiple linear regression) model, in terms of correlation coefficient with the yield statistics. We found that the diurnal temperature range and potential evapotranspiration were the critical factors for rice yield prediction. Our DNN model was also robust to extreme conditions such as drought in 2006 and 2007 in China, which showed its applicability to the future simulation of crop yields under climate change.

본 연구에서는 아스팔트 콘크리트 도로포장의 표면균열 검출을 위해 합성곱 신경망을 이용하였다. 합성곱 신경망의 학습에 사용되는 표면균열 이미지 데이터의 양에 따른 합성곱 신경망의 성능향상 정도를 평가하였다. 사용된 합성곱 신경망의 구조는 5개의 층으로 구성되 어있으며, 3x3 크기의 convolution filter와 2x2 크기의 pooling kernel을 사용하였다. 합성곱 신경망의 학습을 위해서 도로노면 조사 장비를 통해 구축된 국내 도로포장 표면균열 이미지를 활용하였다. 표면균열 이미지 데이터를 학습한 합성곱 신경망 모델의 표면균열 검출 정확도, 정밀도, 재현율, 미검출율, 과검출율을 평가하였다. 가장 많은 양의 데이터를 학습한 합성곱 신경망 모델의 표면균열 검출 정확도, 정밀도, 재현율은 96.6% 이상, 미검출율, 과검출율은 3.4% 이하의 성능을 나타내었다.

Ensemble verification and prediction of low-level wind shear (LLWS) are an important matter for airplane landing and management. In this study, we compared the prediction performance of LLWS forecasts of ensemble mean, multiple regression model and long short-term memory (LSTM), which belong to the family of recurrent neural network based on the grid points over the Jeju area. The prediction skills of methods were compared by mean absolute error. We found that the prediction skills of forecasts of LSTM were better than the bias-corrected forecasts in terms of deterministic prediction.

Though the importance of spectral characteristics of Korean vowels in the hVd syllable has been recognized in the literature, it has never been studied whether static spectral measurements at a (steady-state) central section are enough to characterize Korean vowels in spontaneous speech, or dynamic spectral measurements across the temporal dimension can better characterize vowels. Despite ample reported evidence of the perceptual influence of non-spectral cues on spectral properties of vowels in the literature, no reports have yet been released on the difference in the degree of the perceptual influence of non-spectral cues (e.g., place and manner of the preceding or following phones, F0, speaking rate, prosody, and gender) on spectral properties of vowels. Through Neural Network pattern recognition modeling in a supervised mode, it was found that dynamic spectral models with non-spectral cues better explain vowel perception than static spectral models and furthermore, flanking phone identities, and manner and place of flanking phones are perceptually the most influential while duration, F0 and speaking rate are perceptually far less contributive than argued in the literature.

The optimal grasping point of the object varies depending on the shape of the object, such as the weight, the material, the grasping contact with the robot hand, and the grasping force. In order to derive the optimal grasping points for each object by a three fingered robot hand, optimal point and posture have been derived based on the geometry of the object and the hand using the artificial neural network. The optimal grasping cost function has been derived by constructing the cost function based on the probability density function of the normal distribution. Considering the characteristics of the object and the robot hand, the optimum height and width have been set to grasp the object by the robot hand. The resultant force between the contact area of the robot finger and the object has been estimated from the grasping force of the robot finger and the gravitational force of the object. In addition to these, the geometrical and gravitational center points of the object have been considered in obtaining the optimum grasping position of the robot finger and the object using the artificial neural network. To show the effectiveness of the proposed algorithm, the friction cone for the stable grasping operation has been modeled through the grasping experiments.

This paper presents a 6-DOF relocalization using a 3D laser scanner and a monocular camera. A relocalization problem in robotics is to estimate pose of sensor when a robot revisits the area. A deep convolutional neural network (CNN) is designed to regress 6-DOF sensor pose and trained using both RGB image and 3D point cloud information in end-to-end manner. We generate the new input that consists of RGB and range information. After training step, the relocalization system results in the pose of the sensor corresponding to each input when a new input is received. However, most of cases, mobile robot navigation system has successive sensor measurements. In order to improve the localization performance, the output of CNN is used for measurements of the particle filter that smooth the trajectory. We evaluate our relocalization method on real world datasets using a mobile robot platform.

Ground penetrating radar (GPR) is a typical sensor system for underground objects detection area. The multichannel GPR devices can give more detail and informative three-dimensional (3D) data for classification underground objects. Spatial information of underground objects can be well characterized in the three-dimensional GPR block data which consists of several B-scan and C-scan data. In this article underground object classification method is proposed by using 3D GRP data. Deep learning technique is recently being adopted into this field due to its powerful image classification capacity. The 3D GRP block data is then used to train deep three-dimensional convolution neural network (3D CNN). The proposed method successfully classifies cavity, pipe, manhole and subsoils having small false positive errors. The suggested method is experimentally validated by area data collected on urban roads in Seoul, South Korea.

본 연구에서는 순환신경망을 이용한 댐 유입량 예측모형의 적용성 검토를 목적으로 하고 있으며, 이를 위해 소양강댐 유역 및 충주댐 유역을 대상 으로 그간 댐 운영을 통해 축적된 기상 및 수문 빅데이터를 활용하여 인공신경망 모형과 엘만 순환신경망 모형을 구축하였다. 모형의 학습과 예측 을 위하여 유역별 유입량, 강우량, 기온, 일조시간, 풍속자료가 입력자료로 사용되었고 10일간 일별 댐유입량 자료가 모델의 출력자료로 구조화 하여 학습을 진행한 후 검증을 목적으로 2016년 7월 ~ 2018년 6월까지 2개년에 대한 댐 유입량 예측을 수행하였다. 학습된 모형의 유입량 예측 결과를 비교분석한 결과, 소양강댐 유역에서는 인공신경망 모형과 순환신경망 모형 간 예측성능은 큰 차이를 보이지 않았으며, 충주댐 유역에서는 순환신경망 모형의 예측 결과가 인공신경망 모형에 비해 비교적 우수한 성능을 보임에 따라 엘만 순환신경망을 이용하여 댐 유입량 예측모형을 구축 할 경우 예측성능은 기존의 인공신경망 모형과 비슷하거나 다소 우수할 것으로 판단된다. 또한 엘만 순환신경망은 갈수기 댐 유입량 예측에 있어서 인공신경망에 비해 예측결과의 재현성이 우수한 것으로 나타났으며, 엘만 순환신경망 학습에 있어 다중 은닉층 구조가 단일 은닉층 구조보다 예측 성능 향상에 효과적인 것으로 분석되었다.

3D depth perception has played an important role in robotics, and many sensory methods have also proposed for it. As a photodetector for 3D sensing, single photon avalanche diode (SPAD) is suggested due to sensitivity and accuracy. We have researched for applying a SPAD chip in our fusion system of time-of-fight (ToF) sensor and stereo camera. Our goal is to upsample of SPAD resolution using RGB stereo camera. Currently, we have 64 x 32 resolution SPAD ToF Sensor, even though there are higher resolution depth sensors such as Kinect V2 and Cube-Eye. This may be a weak point of our system, however we exploit this gap using a transition of idea. A convolution neural network (CNN) is designed to upsample our low resolution depth map using the data of the higher resolution depth as label data. Then, the upsampled depth data using CNN and stereo camera depth data are fused using semi-global matching (SGM) algorithm. We proposed simplified fusion method created for the embedded system.

최근의 인공 신경망(Neural Network) 기법은 전통적인 분류 문제와 군집화 문제 해결에서 벗어나 이미지 생성 같은 컨텐츠 생성에서도 좋은 성능을 보이고 있다. 본 연구에서는 차세대 컨텐츠 생성 기법으로 인공신경망을 이용한 이미지 생성기법을 제안한다. 제안하는 인공신경망 모델은 두 개의 이미지를 입력받아서 하나의 이미지에서는 색상을, 다른 이미지에서는 모양을 가져와 새로운 이미지로 조합해낸다. 이 모델은 컨볼루션 인공신경망(Convolutional Neural Network)으로 제작되 었으며 각각 이미지에서 색상과 모양을 추출해내는 두 개의 인코더와 각 인코더의 값을 모두 넘겨 받아 하나의 조합이 되는 이미지를 생성해내는 하나의 디코더로 구성이 되어있다. 본 연구의 성과는 저비용으로 게임 개발 프로세스 상 다양한 2차원 이미지 생성 및 보정 작업에 활용될 수 있다.

This paper presents a convolutional neural network to automatically conduct the peak picking in frequency domain of structural responses. The peaks in frequency domain have a high potential to be the natural frequencies, which are one of the important indicator to be used for structural health monitoring purposes, such as damage detection, cable tension estimation, and finite element model updating. In general, the peaks with the corresponding natural frequencies are manually selected by the users from the frequency domain. Although this previous approach is possible to simply extract the candidate of natural frequencies, it is inappropriate in the practical applications of the long-term monitoring and the implementation for wireless smart sensor. To overcome the drawbacks, this study proposes the convolutional neural network that can automatically identify the peaks with the corresponding natural frequencies from the frequency domain of structural responses.

The talking head (TH) indicates an utterance face animation generated based on text and voice input. In this paper, we propose the generation method of TH with facial expression and intonation by speech input only. The problem of generating TH from speech can be regarded as a regression problem from the acoustic feature sequence to the facial code sequence which is a low dimensional vector representation that can efficiently encode and decode a face image. This regression was modeled by bidirectional RNN and trained by using SAVEE database of the front utterance face animation database as training data. The proposed method is able to generate TH with facial expression and intonation TH by using acoustic features such as MFCC, dynamic elements of MFCC, energy, and F0. According to the experiments, the configuration of the BLSTM layer of the first and second layers of bidirectional RNN was able to predict the face code best. For the evaluation, a questionnaire survey was conducted for 62 persons who watched TH animations, generated by the proposed method and the previous method. As a result, 77% of the respondents answered that the proposed method generated TH, which matches well with the speech.

This paper proposes a convolutional neural network model for distinguishing areas occupied by obstacles from a LiDAR image converted from a 3D point cloud. The channels of a LiDAR image used as input consist of the distances to 3D points, the reflectivities of 3D points, and the heights of 3D points from the ground. The proposed model uses a LiDAR image as an input and outputs a result of a segmented LiDAR image. The proposed model adopts refinement modules with skip connections to segment a LiDAR image. The refinement modules with skip connections in the proposed model make it possible to construct a complex structure with a small number of parameters than a convolutional neural network model with a linear structure. Using the proposed model, it is possible to distinguish areas in a LiDAR image occupied by obstacles such as vehicles, pedestrians, and bicyclists. The proposed model can be applied to recognize surrounding obstacles and to search for safe paths.

하천 관리에 있어 도달시간은 중요한 인자 중의 하나이다. 특히 사회적으로 다양한 하천 활용에 대한 요구가 높아짐에 따라 친수공간으로써 하천에서의 정확한 도달시간 산정은 홍수시 주민 대피 시간 확보 등을 위해서 매우 중요하다. 그러나 과거 도달시간 산정에 대한 연구는 자연 하천의 복합 유역에서의 단일 수문사상에 대하여 연구가 수행되어왔으며, 도심하천의 단일유역을 대상으로 복합 수문 사상에 대한 도달시간 산정방법의 개발은 미흡한 실정이다. 따라서 최근 집중호우에 의하여 빈번한 침수 피해가 발생된 부산광역시 대표 도심하천인 온천천 유역에 대하여 과거 10년(2006~2015년) 동안의 강우-유출량 자료를 이용하여 도달시간을 산정하였고, Matlab 기반의 인공신경망 기법을 이용하여 신뢰성을 검토하였다. 12시간 이상 무강우를 기준으로 총 254개의 강우 사상을 분리하였고, 이를 바탕으로 총 강우량, 총 유출량, 첨두 강우량/총 강우량, 첨두 유출량/총 유출량, 지체시간, 도달시간 등 총 6개의 변수를 산정하여 인공신경망 모형의 훈련 및 검증에 활용하였다. 그 결과 훈련에 과 예측 및 검증에 활용된 입력 변수의 상관관계는 각 각 0.807 및 0.728로 나타났으며, 연구결과를 바탕으로 도심하천의 도달시간 산정결과의 신뢰성 분석에 이를 활용할 수 있을 것으로 판단된다.

본 연구는 인공신경망을 이용해 철골모멘트골조의 접합부 손상을 예측하는 기법을 제안한다. 인공신경망의 입력층에는 기둥 부재 의 휨모멘트, 고유진동수, 모드형상 정보가 사용되며, 출력층에는 구조물 접합부의 회전강성 손상지표가 사용한다. 손상지표는 각 접합부의 손 상정도를 의미한다. 5층 철골모멘트골조 예제의 수치해석을 통해 훈련 및 검증용 데이터를 생성한다. 총 829가지의 손상 시나리오가 고려된다. 시뮬레이션은 OpenSees를 이용해 반복 실행하여 데이터를 얻도록 하였으며, 훈련용 데이터를 생성할 때 회전 강성의 손상은 1.0, 0.75, 0.5 등 세 가지 중 하나의 값을 가지도록 하였다. 예제 검증을 통해 제시하는 기법은 손상 위치 및 수준을 정확하게 예측하는 것으로 나타났다. 제시하는 기법은 손상지표, 1차, 2차 고유진동수 및 모드형상 등에 대해 매우 유사한 결과를 제시하는 것으로 확인되었다.

In this paper, we propose a jellyfish distribution recognition and monitoring system using a UAV (unmanned aerial vehicle). The UAV was designed to satisfy the requirements for flight in ocean environment. The target jellyfish, Aurelia aurita, is recognized through convolutional neural network and its distribution is calculated. The modified deep neural network architecture has been developed to have reliable recognition accuracy and fast operation speed. Recognition speed is about 400 times faster than GoogLeNet by using a lightweight network architecture. We also introduce the method for selecting candidates to be used as inputs to the proposed network. The recognition accuracy of the jellyfish is improved by removing the probability value of the meaningless class among the probability vectors of the evaluated input image and re-evaluating it by normalization. The jellyfish distribution is calculated based on the unit jellyfish image recognized. The distribution level is defined by using the novelty concept of the distribution map buffer.