The input vector space optimization for LSTM deep learning model in real-time prediction of ship motions

Liu, Yucheng; Duan, Wenyang; Huang, Li‐Min; Duan, Shiliang; Ma, Xuefei

doi:10.1016/j.oceaneng.2020.107681

Cited by 66 publications

(18 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of recurrent neural networks, LSTM-based models presented good performance in time series classification tasks and prediction tasks [40]. The LSTM network is useful in solving non-linear problems due to its non-linear processing capacity [41].…”

Section: Mlp and Recurrent Networkmentioning

confidence: 99%

Comparing LSTM and GRU Models to Predict the Condition of a Pulp Paper Press

et al. 2021

View full text Add to dashboard Cite

The accuracy of a predictive system is critical for predictive maintenance and to support the right decisions at the right times. Statistical models, such as ARIMA and SARIMA, are unable to describe the stochastic nature of the data. Neural networks, such as long short-term memory (LSTM) and the gated recurrent unit (GRU), are good predictors for univariate and multivariate data. The present paper describes a case study where the performances of long short-term memory and gated recurrent units are compared, based on different hyperparameters. In general, gated recurrent units exhibit better performance, based on a case study on pulp paper presses. The final result demonstrates that, to maximize the equipment availability, gated recurrent units, as demonstrated in the paper, are the best options.

show abstract

Section: Mlp and Recurrent Networkmentioning

confidence: 99%

Comparing LSTM and GRU Models to Predict the Condition of a Pulp Paper Press

et al. 2021

View full text Add to dashboard Cite

show abstract

“…It means that RNNs equally update information in the sequence, even when the time interval is not equal. Other researchers applied LSTM in the vessel domain, but they did not consider irregularity of time [9,11]. This implies the necessity of a sequence model addressing irregular time intervals.…”

Section: Sequence Modelmentioning

confidence: 99%

“…Although the authors of [7] adopted MLP, they addressed the importance of sequential property in ship data. On the other hand, the authors of [9][10][11] applied Long Short-Term Memory (LSTM) [12] to consider sequential ship data. Among them, the authors of [9,10] proved that LSTM outperformed MLP and other machine learning models which could not reflect sequential information.…”

Section: Introductionmentioning

confidence: 99%

Time-Aware and Feature Similarity Self-Attention in Vessel Fuel Consumption Prediction

Park

Lee

Park³

et al. 2021

Applied Sciences

View full text Add to dashboard Cite

An accurate vessel fuel consumption prediction is essential for constructing a ship route network and vessel management, leading to efficient sailings. Besides, ship data from monitoring and sensing systems accelerate fuel consumption prediction research. However, the ship data consist of three properties: sequential, irregular time interval, and feature importance, making the predicting problem challenging. In this paper, we propose Time-aware Attention (TA) and Feature-similarity Attention (FA) applied to bi-directional Long Short-Term Memory (LSTM). TA acquires time importance by nonlinear function from irregular time intervals in each sequence and emphasizes data depending on the importance. FA emphasizes data based on similarities of features in the sequence by estimating feature importance with learnable parameters. Finally, we propose the ensemble model of TA and FA-based BiLSTM. The ensemble model, which consists of fully connected layers, is capable of simultaneously capturing different properties of ship data. The experimental results on ship data showed that the proposed model improves the performance in predicting fuel consumption. In addition to model performance, visualization results of attention maps and feature importance help to understand data properties and model characteristics.

show abstract

“…The data-driven methods learn patterns from historical observations and further use the learned patterns to predict future vessels motion. The common strategies are real-time prediction [8]- [10], and short-term prediction [11]. The machine learning-based prediction models were used for the real-time vessels motion prediction to their capability in nonlinearity processing.…”

Section: Introductionmentioning

confidence: 99%

Short-Term Prediction in Vessel Heave Motion Based on Improved LSTM Model

et al. 2021

View full text Add to dashboard Cite

In order to solve the problem of control performance degradation caused by time delay in wave compensation control system, predicting vessel heave motion can be the input vector of the control system to alleviate time delay problem. The vessel heave motion belongs to the problem of time series, this paper proposes an improved Long Short-Term Memory (LSTM) model with a random deactivation layer (dropout), which can deal with the time series problem very well. In order to obtain the vessel heave motion, this paper establishes a wave model suitable for marine operation, and solves the vessel heave motion through the mathematical model of vessel motion. Finally, the paper predicts the vessel heave motion in a short predicted time series. In the process of obtaining the prediction effect of vessel heave motion, the Back Propagation (BP) neural network and the standard LSTM neural network are used to compare with the improved LSTM neural network. While the predicted time series is 0.1 s at sea state 3, the mean absolute percentage (MAPE) errors of BP neural network in the prediction of vessel heave motion is 1.06 × 10 −2 %, the standard LSTM in the prediction of heave motion is 1.43×10 −4 %, the improved LSTM in the prediction of heave motion is 7.51 × 10 −6 %. The improved LSTM improves MAPE by 1.05 × 10 −2 % compared with the BP and 1.42 × 10 −4 % compared with the standard LSTM. The prediction results show that the improved LSTM has a strong prediction capability with not easily overfitted in vessel heave motion prediction. The results show that the improved LSTM provides a new idea for vessel motion prediction and solves the problem of time delay, which is useful for the study of stability in marine operations.INDEX TERMS Short-term prediction, improved long short-term memory (LSTM), vessel heave motion, time delay.

show abstract

The input vector space optimization for LSTM deep learning model in real-time prediction of ship motions

Cited by 66 publications

References 9 publications

Comparing LSTM and GRU Models to Predict the Condition of a Pulp Paper Press

Comparing LSTM and GRU Models to Predict the Condition of a Pulp Paper Press

Time-Aware and Feature Similarity Self-Attention in Vessel Fuel Consumption Prediction

Short-Term Prediction in Vessel Heave Motion Based on Improved LSTM Model

Contact Info

Product

Resources

About