Long short term memory 모델을 이용한 시계열 수중 소음 데이터 예측

이혜선,홍우영,김국현,이근화 한국음향학회 2023 韓國音響學會誌 Vol.42 No.4

In this paper, a time series machine learning model, Long Short Term Memory (LSTM), is applied into the bubble flow noise data and the underwater projectile launch noise data to predict missing values of time-series underwater noise data. The former is mixed with bubble noise, flow noise, and fluid-induced interaction noise measured in a pipe and can be classified into three types. The latter is the noise generated when an underwater projectile is ejected from a launch tube and has a characteristic of instantaenous noise. For such types of noise, a data-driven model can be more useful than an analytical model. We constructed an LSTM model with given data and evaluated the model’s performance based on the number of hidden units, the number of input sequences, and the decimation factor of signal. It is shown that the optimal LSTM model works well for new data of the same type. 본 논문에서는 일부 소음 데이터만 알고 있을 때 결손된 데이터를 예측할 목적으로 수조에서 측정된 기포유동소음 데이터와 수중 운동체 발사 소음 데이터를 시계열 기계학습 모델인 Long Short Term Memory(LSTM)에 적용해보았다. 기포유동소음 데이터는 파이프에서 측정된 소음으로 기포소음, 유동소음, 유체기인소음이 혼합되어 있으며유형별로 3가지로 분류할 수 있다. 수중 운동체 발사소음은 모형 발사튜브에서 수중 운동체가 사출될 때 발생하는 소음으로 순간소음이며 발사 이벤트마다 불규칙하게 변한다. 이러한 종류의 소음 생성을 위해서는 해석적인 모델보다는데이터 기반 모델이 유용할 수 있다. 본 연구에서는 LSTM을 데이터 기반 모델을 만들었다. 모델에 영향을 주는 LSTM 의 은닉유닛의 개수, 입력시퀸스의 개수, 데시메이션 인자에 따른 모델의 성능을 확인하고 최적의 LSTM 모델을 구성했다. 같은 유형은 새로운 데이터에 대해서도 잘 동작하는 것을 보였다.

딥러닝 기반 LSTM 모형을 이용한 항적 추적성능 향상에 관한 연구

황진하(Jin-Ha Hwang),이종민(Jong-Min Lee) 한국정보통신학회 2021 한국정보통신학회 종합학술대회 논문집 Vol.25 No.1

항적추적 기술에 딥러닝 기반 LSTM(Long Short-Term Memory) 모델을 적용하는 연구로서 기존의 항적추적기술의 경우, 항공기의 등속, 등가속, 급기동, 선회(3D) 비행 등 비행 특성에 따른 칼만 필터 기반의 LMIPDA를 활용한 실시간 항적 추적 시 등속, 등가속, 급기동, 선회(3D) 비행 가중치가 자동으로 변경된다. 이러한 과정에서 등속 비행 중 급기동 비행과 같이 비행 특성이 변경될 때, 항적 손실 및 항적 추적 성능이 하락하여 비행 특성 가중치 변경성능을 향상시킬 필요성이 있다. 본 연구는 레이더의 오차 모델이 적용된 시뮬레이터의 Plot과 표적을 딥러닝 기반 LSTM(Long Short-Term Memory) 모델을 적용하여 학습시키고, 칼만 필터를 활용한 항적추적 결과와 딥러닝 기반 LSTM(Long Short-Term Memory) 모델을 적용한 항적추적결과를 비교함으로써 미리 비행 특성의 변경과정을 예측하여 등속, 등가속, 급기동, 선회(3D) 비행 가중치변경을 신속하게 함으로써 항적추적성능을 향상하기 위한 연구이다. This study applies a deep learning-based long short-term memory(LSTM) model to track tracking technology. In the case of existing track tracking technology, the weight of constant velocity, constant acceleration, stiff turn, and circular(3D) flight is automatically changed when tracking track in real time using LMIPDA based on Kalman filter according to flight characteristics of an aircraft such as constant velocity, constant acceleration, stiff turn, and circular(3D) flight. In this process, it is necessary to improve performance of changing flight characteristic weight, because changing flight characteristics such as stiff turn flight during constant velocity flight could incur the loss of track and decreasing of the tracking performance. This study is for improving track tracking performance by predicting the change of flight characteristics in advance and changing flight characteristic weigh rapidly. To get this result, this study makes deep learning-based Long Short-Term Memory(LSTM) model study the plot and target of simulator applied with radar error model, and compares the flight tracking results of using Kalman filter with those of deep learning-based Long Short-Term memory(LSTM) model.

RCS Estimation using LSTM at High Frequency

Geumbi Park(박금비),Jinhwan Koh(고진환) 한국산학기술학회 2022 한국산학기술학회논문지 Vol.23 No.11

RCS 측정은 전자파의 산란과 반사가 필수적인 통신 시스템, 안테나 시스템, 항공기 설계에 도움이 되는 요소이다. 항공기나 선박 등 대형 물체에서 고주파 RCS 데이터를 얻기위해서는 측정 시간과 비용이 많이 들게된다. 이에 본 논문에서는 위와 같은 문제를 해소하고, RCS 측정 효율을 높이기 위하여 AI모델을 도입한다. AI 모델 중 Long Short Term Memory(LSTM)은 장기의존성 문제를 해결하는 장점을 가진다. 따라서 LSTM의 방법을 제안하며, LSTM의 모델에서 정확도를 더 높이고 시간을 단축하고자 병렬 LSTM모델을 제안한다. 컴퓨터 시뮬레이션 CST를 이용하여 RCS를 측정하고, Matlab을 사용하여 CST로 측정된 데이터 중 저주파대역을 데이터를 학습한 후, 고주파대역의 RCS를 LSTM와 병렬 LSTM 모델을 이용하여 추정하였다. 이후 LSTM 모델과 병렬 LSTM 모델의 추정값과 CST로 측정한 값을 비교하여, 허용 범위의 오차 값 이내의 결과를 통해 높은 정확도를 확인하였다. 결과는 CST로 측정하였을 때보다 LSTM과 병렬 LSTM을 사용하였을 때 시간이 크게 단축되었음을 보여주었다. RCS measurements are a helpful factor in the design of communication systems, antenna systems, and aircraft, where scattering and reflection of electromagnetic waves are essential. High-frequency RCS measurement time and expenses are relatively high in large objects, such as aircraft and ships. This paper introduces an AI model to solve the above problems and increase the RCS measurement efficiency. Among AI models, Long Short Term Memory (LSTM) has the advantage of solving the long-term dependence problem. Therefore, a method of LSTM and a parallel LSTM model are proposed to increase the accuracy further and reduce time in the model of LSTM. RCS was measured using a computer simulation CST, and low-frequency band data among CST-measured data was learned using Matlab. The RCS of the high-frequency band was estimated using LSTM and the parallel LSTM model. The estimated value of the LSTM model and the value measured by CST were compared with the estimated value of the parallel LSTM model. The high accuracy was confirmed through the results within the error value of the allowable range. In addition, the time was reduced significantly using LSTM and parallel LSTM than with CST.

Dynamical prediction of two meteorological factors using the deep neural network and the long short-term memory (ΙΙ)

Shin Ki-Hong,Jung Jae-Won,Chang Ki-Ho,Kim Kyungsik,Jung Woon-Seon,Lee Dong-In,You Cheol-Hwan 한국물리학회 2022 THE JOURNAL OF THE KOREAN PHYSICAL SOCIETY Vol.80 No.12

This paper presents the predictive accuracy using two-variate meteorological factors, average temperature and average humidity, in neural network algorithms. We analyze result in fve learning architectures such as the traditional artifcial neural network, deep neural network, and extreme learning machine, long short-term memory, and long-short-term memory with peephole connections, after manipulating the computer simulation. Our neural network modes are trained on the daily time-series dataset during 7 years (from 2014 to 2020). From the trained results for 2500, 5000, and 7500 epochs, we obtain the predicted accuracies of the meteorological factors produced from outputs in ten metropolitan cities (Seoul, Daejeon, Daegu, Busan, Incheon, Gwangju, Pohang, Mokpo, Tongyeong, and Jeonju). The error statistics is found from the result of outputs, and we compare these values to each other after the manipulation of fve neural networks. As using the long-shortterm memory model in testing 1 (the average temperature predicted from the input layer with six input nodes), Tonyeong has the lowest root-mean-squared error (RMSE) value of 0.866 (%) in summer from the computer simulation to predict the temperature. To predict the humidity, the RMSE is shown the lowest value of 5.732 (%), when using the long short-term memory model in summer in Mokpo in testing 2 (the average humidity predicted from the input layer with six input nodes). Particularly, the long short-term memory model is found to be more accurate in forecasting daily levels than other neural network models in temperature and humidity forecastings. Our result may provide a computer simulation basis for the necessity of exploring and developing a novel neural network evaluation method in the future.

LSTM을 활용한 불법주정차 시공간 예측 모델링: 서울시 민원신고 데이터를 중심으로

김동은,강영옥 대한공간정보학회 2019 대한공간정보학회지 Vol.27 No.3

This study aims to predict the number of illegal parking using data on complaints of illegal parking within Seoul. We made the prediction models using long short-term memory(LSTM) which has shown high performance in the field of the series prediction recently. Because setting the time unit is important for time series prediction, prediction models in the previous researches include consideration of how to set the time unit. In this study, the prediction models were made with consideration of not only time unit but also space unit. The time unit of prediction was set up for 24 hours per month, and the space unit was designated as district ‘gu’, land-use type, and road and off the road. As a result, it was confirmed that the prediction models performed well overall when dividing the spatial units by land-use type and road and off the road, and that the prediction models performed poorly when divided into district ‘gu’. 본 연구의 목적은 서울시 내에서 발생한 불법주정차 민원신고 데이터를 활용하여 불법주정차 발생의 시공간 예측모델을 구축하는 것이다. 예측모델은 최근 시계열 예측 분야에서 높은 성능을 보이고 있는 long short-term memory (LSTM)을 활용하여 생성하였다. LSTM을 활용한 시계열 예측 시에는 시간단위 설정이 중요하기 때문에 기존 연구에서의 예측모델은 시간단위를 어떻게 설정할 것인가에 집중하고 있다. 본 연구에서는 시간단위 뿐 아니라 공간단위 설정에 따른 문제점을 분석하고, 실험을 통해 최적의 예측 공간단위를 찾고자 하였다. 이를 위해 예측의 시간단위는 월별 24시간을 기준으로 하고, 공간단위는 자치구, 토지이용유형, 도로 및 도로 외, 그리고 토지이용유형별 도로와 도로가 아닌 지역으로 구분한 공간 단위별로 분석을 수행하였다. 그 결과 토지이용유형, 도로 및 도로 외로 공간단위를 구분하였을 때 예측모델들이 전반적으로 좋은 성능을 보였으며, 자치구로 구분하였을 때 좋지 않은 성능을 보이는 것을 확인하였다.

Long Short-Term Memory를 활용한 건화물운임지수 예측

한민수,유성진 한국품질경영학회 2019 품질경영학회지 Vol.47 No.3

Purpose: The purpose of this study is to overcome limitations of conventional studies that to predict Baltic Dry Index (BDI). The study proposed applications of Artificial Neural Network (ANN) named Long Short-Term Memory (LSTM) to predict BDI. Methods: The BDI time-series prediction was carried out through eight variables related to the dry bulk market. The prediction was conducted in two steps. First, identifying the goodness of fitness for the BDI time-series of specific ANN models and determining the network structures to be used in the next step. While using ANN's generalization capability, the structures determined in the previous steps were used in the empirical prediction step, and the sliding-window method was applied to make a daily (one-day ahead) prediction. Results: At the empirical prediction step, it was possible to predict variable (BDI time series) at point of time  by 8 variables (related to the dry bulk market) of  at point of time    . LSTM, known to be good at learning over a long period of time, showed the best performance with higher predictive accuracy compared to Multi-Layer Perceptron (MLP) and Recurrent Neural Network (RNN). Conclusion: Applying this study to real business would require long-term predictions by applying more detailed forecasting techniques. I hope that the research can provide a point of reference in the dry bulk market, and furthermore in the decision-making and investment in the future of the shipping business as a whole.

Long Short-Term Memory를 이용한 부산항 조위 예측

김해림,전용호,박재형,윤한삼 해양환경안전학회 2022 해양환경안전학회지 Vol.28 No.4

This study developed a Recurrent Neural Network model implemented through Long Short-Term Memory (LSTM) that generates long-term tidal level data at Busan Port using tide observation data. The tide levels in Busan Port were predicted by the Korea Hydrographic and Oceanographic Administration (KHOA) using the tide data observed at Busan New Port and Tongyeong as model input data. The model was trained for one month in January 2019, and subsequently, the accuracy was calculated for one year from February 2019 to January 2020. The constructed model showed the highest performance with a correlation coefficient of 0.997 and a root mean squared error of 2.69 cm when the tide time series of Busan New Port and Tongyeong were inputted together. The study’s finding reveal that long-term tidal level data prediction of an arbitrary port is possible using the deep learning recurrent neural network model. 본 연구는 조위 관측자료를 이용하여 부산항에서의 장기 조위 자료를 생성하는 Long Short-Term Memory (LSTM)으로 구현된 순환신경망 모델을 개발하였다. 국립해양조사원의 부산 신항과 통영에서 관측된 조위 자료를 모델 입력 자료로 사용하여 부산항의 조위를 예측하였다. 모델에 대하여 2019년 1월 한 달의 학습을 수행하였으며, 이후 2019년 2월에서 2020년 1월까지 1년에 대하여 정확도를 계산하였다. 구축된 모델은 부산 신항과 통영의 조위 시계열을 함께 입력한 경우에 상관계수 0.997 및 평균 제곱근 오차 2.69m로 가장 성능이 높았다. 본 연구 결과를 바탕으로 딥러닝 순환신경망 모델을 이용하여 임의 항만의 장기 조위 자료 예측이 가능함을 알 수 있었다.

CMIP5 기반 하천유량 예측을 위한 딥러닝 LSTM 모형의 최적 학습기간 산정

천범석 ( Chun Beom-Seok ),이태화 ( Lee Tae-hwa ),김상우 ( Kim Sang-woo ),임경재 ( Lim Kyoung-jae ),정영훈 ( Jung Young-hun ),도종원 ( Do Jong-won ),신용철 ( Shin Yong-chul ) 한국농공학회 2022 한국농공학회논문집 Vol.64 No.1

In this study, we suggested the optimal training period for predicting the streamflow using the LSTM (Long Short-Term Memory) model based on the deep learning and CMIP5 (The fifth phase of the Couple Model Intercomparison Project) future climate scenarios. To validate the model performance of LSTM, the Jinan-gun (Seongsan-ri) site was selected in this study. We comfirmed that the LSTM-based streamflow was highly comparable to the measurements during the calibration (2000 to 2002/2014 to 2015) and validation (2003 to 2005/2016 to 2017) periods. Additionally, we compared the LSTM-based streamflow to the SWAT-based output during the calibration (2000∼2015) and validation (2016∼2019) periods. The results supported that the LSTM model also performed well in simulating streamflow during the long-term period, although small uncertainties exist. Then the SWAT-based daily streamflow was forecasted using the CMIP5 climate scenario forcing data in 2011∼2100. We tested and determined the optimal training period for the LSTM model by comparing the LSTM-/SWAT-based streamflow with various scenarios. Note that the SWAT-based streamflow values were assumed as the observation because of no measurements in future (2011∼2100). Our results showed that the LSTM-based streamflow was similar to the SWAT-based streamflow when the training data over the 30 years were used. These findings indicated that training periods more than 30 years were required to obtain LSTM-based reliable streamflow forecasts using climate change scenarios.

다변량 LSTM을 이용한 극저온 액화가스 저장탱크의 슬로싱에 의한 첨두 압력 예측

김재원(Jaewon Kim),정소명(Somyung Chung),전규목(Gyumok Jeon),박종천(Jongchun Park) 한국해양환경·에너지학회 2021 한국해양환경·에너지학회 학술대회논문집 Vol.2021 No.10

본 연구에서는 극저온 액화가스 저장 탱크 내부에서 발생하는 슬로싱에 의한 탱크 내부의 첨두 압력을 계산하기 위한 전산유체 시뮬레이션과, 시뮬레이션을 통한 계산값 이후의 첨두 압력 시계열 결과값 예측을 위해 딥러닝 방식의 순환 신경망 구조(Recurrent Neural Network, RNN) 중 하나인 LSTM(Long Short-Term Memory)를 이용하는 방법론을 제시한다. 극저온 액화가스 저장탱크의 첨두 압력 계산을 위한 상용 유동해석 시뮬레이션 프로그램으로 STAR-CCM+(ver. 15.02.007)을 사용하였으며, 계산된 첨두 압력 시계열을 Python 기반 Tensorflow 라이브러리의 LSTM 모델을 통해 이후 시계열 값을 예측하였다. 이 때 저장탱크 내 첨두 압력 시계열에 직접적인 영향을 미치는 것으로 판단되는 변수를 다변량으로 LSTM 모델에 포함시켜 계산을 진행하였으며, 이를 시뮬레이션 계산 결과와 비교하여 LSTM을 통한 예측값의 유효성을 판단하였다. This study presents a methodology of using LSTM(Long Short-Term Memory), one of the deep learning Recurrent Neural Networks(RNN), to calculate peak pressure inside the cryogenic storage tank by sloshing generated inside the tank and predict peak pressure time series after calculation. STAR-CCM+(ver.15.02.007) was used in this study for calculation of peak pressure in the tank, and LSTM module in Tensorflow(ver.2.4.0) based Python was used in this study for prediction of time series after calculations. Then the calculation was performed by including variables that were determined to have a direct effect on the peak pressure time series in the storage tank in a Multivariate LSTM, and the validity of the predicted value through LSTM was determined by comparing them with the simulation and calculation results.

Long Short Term Memory based Political Polarity Analysis in Cyber Public Sphere

Hyeon Kang,Dae-Ki Kang 국제문화기술진흥원 2017 International Journal of Advanced Culture Technolo Vol.5 No.4

In this paper, we applied long short term memory(LSTM) for classifying political polarity in cyber public sphere. The data collected from the cyber public sphere is transformed into word corpus data through word embedding. Based on this word corpus data, we train recurrent neural network (RNN) which is connected by LSTM’s. Softmax function is applied at the output of the RNN. We conducted our proposed system to obtain experimental results, and we will enhance our proposed system by refining LSTM in our system.