Indonesia is currently focusing on its big goal to become The World’s Maritime Axis. For this reason, several sectors such as the infrastructure of the port, the development of the fishing, and tourism industry should be improved. The use of accurate tides level data is indispensable to support these developments. However, the number of instruments to observe tides data is limited compared to the covered area since Indonesia has the third longest coastline in the world. Recently, the frequent use of Artificial Intelligence (AI) has also offered an alternative solution to provide prediction data, including tides level data. Thereby, Artificial Neural Networks (ANN) as the subfield of AI is then chosen to make a prediction of tides level data. The type of ANN used in this study is two-layer Feed Forward Neural Network (FFNN). The previous observed tides data using atmospheric data (temperature and pressure) and moon position as the features are used to train the network. In order to evaluate the performance of ANN model, the result of the prediction is then compared to the observed tides level data using Automatic Weather Station (AWS). The result shows that the predicted tide level data has a strong correlation with the observed data with coefficient correlation of 0.9238. Furthermore, Root Mean Square Error (RMSE) as the statistics parameters to evaluate the performance of ANN model is found to be low around 0.077 meters. This preliminary result suggests that the FFNN has a good performance in predicting tides level data and therefore can be applied to provide tides level data on a larger scale in Indonesia.

The Implementation of Artificial Neural Network (ANN) in the Prediction of Tides Level Data in Indonesia

Images taken with different sensors and transmitted through different channels can be noisy. In such conditions, the image most often suffers from random-valued impulse noise. Denoising an image is an important part of image preprocessing before recognition by a neural network. The accuracy of image recognition by a neural network directly depends on the intensity of image noise. This paper presents a three-stage image cleaning and recognition system, which includes a developed detector of pulsed noisy pixels, a filter for cleaning found noisy pixels based on an adaptive median, and a neural network program for recognizing cleaned images. It was noted that at low noise intensities, cleaning is practically not required, but noise with an intensity of more than 10% can seriously damage the image and reduce recognition accuracy. As a training base for noise, cleaning, and recognition, the CIFAR10 digital image database was used, consisting of 60,000 images belonging to 10 classes. The results show that the proposed neural network recognition system for images affected by to random-valued impulse noise effectively finds and corrects damaged pixels. This helped to increase the accuracy of image recognition compared to existing methods for cleaning random-valued impulse noise.

/pdf/neural-network-system-for-recognizing-images-affected-by-urer5cpl.pdf

Neural Network System for Recognizing Images Affected by Random-Valued Impulse Noise

Indonesia is currently focusing on its big goal to become The World’s Maritime Axis. For this reason, several sectors such as the infrastructure of the port, the development of the fishing, and tourism industry should be improved. The use of accurate tides level data is indispensable to support these developments. However, the number of instruments to observe tides data is limited compared to the covered area since Indonesia has the third longest coastline in the world. Recently, the frequent use of Artificial Intelligence (AI) has also offered an alternative solution to provide prediction data, including tides level data. Thereby, Artificial Neural Networks (ANN) as the subfield of AI is then chosen to make a prediction of tides level data. The type of ANN used in this study is two-layer Feed Forward Neural Network (FFNN). The previous observed tides data using atmospheric data (temperature and pressure) and moon position as the features are used to train the network. In order to evaluate the performance of ANN model, the result of the prediction is then compared to the observed tides level data using Automatic Weather Station (AWS). The result shows that the predicted tide level data has a strong correlation with the observed data with coefficient correlation of 0.9238. Furthermore, Root Mean Square Error (RMSE) as the statistics parameters to evaluate the performance of ANN model is found to be low around 0.077 meters. This preliminary result suggests that the FFNN has a good performance in predicting tides level data and therefore can be applied to provide tides level data on a larger scale in Indonesia. 

Tide level forecasting is one of major components in oceanography and marine research. The future prediction of tides is an important issue for human and environment. This study proposes new tide level prediction methods with multiple input multiple output (MIMO) approach based on nonlinear autoregressive with exogenous inputs (NARX) model and three types of recurrent neural networks (RNNs), namely, long short-term memory (LSTM), bi-directional long short-term memory (biLSTM) and gated recurrent unit (GRU) networks. We develop three types of NARX models: NARX-LSTM, NARX-biLSTM and NARX-GRU. We evaluate the three models and compare their performances with LSTM, biLSTM and GRU network models. It turns out that the proposed MIMO NARX based RNN models, especially NARX-LSTM and NARX-GRU have better performance in comparison with their RNN counterparts (i.e., LSTM, biLSTM, and GRU). The NARX-GRU model shows the best performance for one step and three step ahead prediction in terms of RMSE in each location and average RMSE, while NARX-LSTM performs better for two steps ahead prediction in terms average RMSE. The results also show that the mixture of RNN and NARX improves the performance of RNN model for all the four tasks at four locations in the UK (Wick, Workington, Ullapol and New Heaven) and for the three time horizons (i.e., 1, 2 and 3 steps ahead).

Tide Level Prediction Using NARX-based Recurrent Neural Networks

In recent years, most research on typhoons in the Taiwan Strait and its adjacent waters has focused on simulating typhoon waves under the influence of wind fields. In order to study the influence of tidal level changes on typhoon waves, a numerical model was established in the Taiwan Strait based on the third-generation ocean wave model SWAN. The simulation results of the tide level during the corresponding typhoon landing time were incorporated into the model to optimize its performance. Subsequently, the wave height of the typhoon landing at the lowest tide level was compared with that at the highest tide level. This comparison serves as a reference and warning for ocean engineering, highlighting the hazards of the typhoon landing at high tide. The simulation results were verified and analyzed using the measured data of significant wave heights and wind speeds when typhoons Mekkhala (2006) and Maria (0607) approached. The results show that after optimization, the relative error of the significant wave peak is reduced. Furthermore, there has been a decrease in the maximum wind speed, bringing it closer to the measured value. These improvements signify enhanced model accuracy. The tide level has a great influence on the typhoon wave, and the tide level height at the time of the typhoon landing is positively correlated with the significant wave height of the waves generated by the typhoon. When the typhoon’s landing time coincides with the high tide level, the resulting waves are significantly higher, reaching up to 0.71 m. This has a substantial impact on the safety of marine structures, particularly breakwaters.

https://www.mdpi.com/2073-4441/15/10/1807/pdf?version=1683626880

Effect of Tide Level Change on Typhoon Waves in the Taiwan Strait and Its Adjacent Waters

Tide variations are affected not only by periodic movement of celestial bodies but also by time-varying interference from the external environment. To improve the accuracy of tide prediction, a modular tide level prediction model (HA-NARX) is proposed. This model divides tide data into two parts: astronomical tide data affected by celestial tide-generating forces and nonastronomical tide data affected by various environmental factors. Final tide prediction results are obtained using a nonlinear autoregressive exogenous model (NARX) neural network combined with harmonic analysis (HA) data. To verify the feasibility of the model, tide data under different climatic and geographical conditions are used to simulate the prediction of tide levels, and the results are compared with those of traditional HA, the genetic algorithm-back propagation (GA-BP) neural network and the wavelet neural network (WNN). The results show that the greater the influence of meteorological factors on tides, the more obvious is the improvement in accuracy and stability of HA-NARX prediction results compared to traditional models, with the highest prediction accuracy improvement of 234%. The proposed model not only has a simple structure but can also effectively improve the stability and accuracy of tide prediction.

/pdf/a-modular-tide-level-prediction-method-based-on-a-narx-8fmm9ex2cg.pdf

Wenyu Lyu

Papers

A Modular Tide Level Prediction Method Based on a NARX Neural Network