Multi-sensor Fusion with Hybrid Deep Learning Convolutional Neural Network-Long Short Term Memory for Real-time Welding Activity Recognition and Occupational Safety Compliance

The prediction of land subsidence is a crucial step for early warning of urban infrastructure damage and timely remedy. However, the performance of most mathematical and empirical prediction models is often compromised by their large number of parameters, complex operational processes and sparsely measured values. Currently, the traditional neural network models are popular and effective, but they cannot accurately discover the characteristic changes of time series data. In this paper, a long short-term memory (LSTM) neural network was proposed to predict the land subsidence of time series Interferometric Synthetic Aperture Radar (InSAR). First, the Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) technique was utilized to monitor the time series land subsidence at Beijing Capital International Airport (BCIA) from 2005 to 2010 based on ENVISAT ASAR images with a descending orbit. The results were compared with the existing results to verify the reliability and then used to analyse the temporal and spatial characteristics of the time series land subsidence of the BCIA. Based on the time series InSAR deformation data, the LSTM neural network was used to establish the prediction model of time series InSAR, and the results were compared with those of the Multi-Layer Perceptron (MLP) and Recurrent Neural Network (RNN). The comparison results showed that the LSTM neural network was more accurate than the MLP and RNN on the point scale (the root mean square error was 4.60 mm and the mean absolute error was 3.18 mm), the correlation coefficients between the prediction results of the LSTM neural network and the real InSAR measurement results in 2007 and 2008 were 0.93 mm and 0.96 mm, respectively, indicating that LSTM neural network had better prediction performance. Eventually, based on the land subsidence data of time series InSAR from 2006 to 2010, the LSTM neural network was applied to predict the BCIA time series land subsidence in 2011. The results predicted that cumulative subsidence in September 2011 would reach a maximum of 350 mm. Therefore, the LSTM neural network is a potentially effective prediction method, which can replace numerical or empirical models in the absence of detailed hydrogeological data. Moreover, its prediction results can be used to assist decision-making, early warning and hazard relief.

Oil and gas consumption for power generation has caused irreversible damage to humanity. To address the attendant effects of fossil fuel utilization, renewable energy is a good alternative. International organizations give support to countries in their transition to a green energy future. This implies that the use of renewable energy is widely supported. It is therefore recommended to utilize renewable energy as it is environmentally friendly. One such type of renewables is water energy. Water cycle has streamflow $$\left({f_{s}} \right)$$ f s as its central component. Having reliable information about future $$f_{s}$$ f s data is essential in hydrological research, as it can help water managers to plan hydropower generation. To also ensure preparedness and mitigation of floods and drought as well as hydropower production planning and management, precise $$f_{s}$$ f s prediction is considered essential. Several modelling methods have been used lately to forecast $$f_{s}$$ f s , namely: physical methods, data-driven approaches such as shallow artificial neural networks (ANNs), and hybrid techniques. Nevertheless, they may not approximate complex relationships as accurate as deep learning techniques. In this study, an innovative deep learning technique based on long short-term memory (LSTM) neural network adapted with data preprocessing algorithm (DPA) is proposed for seasonal $$f_{s}$$ f s forecasting. Considering recent studies on $$f_{s}$$ f s forecasting, one can avow that researchers have been able to employ lag value predictors for future $$f_{s}$$ f s extrapolation, although deep learning techniques can offer good potentials for $$f_{s}$$ f s prediction with complex physical relationship. However, to the best knowledge of the authors, only very few studies have applied LSTM neural network for streamflow forecasting. In addition, there have been attempts to estimate river $$f_{s}$$ f s in Nigeria using some traditional methods though, but the effect of seasonal variation on $$f_{s}$$ f s forecasting has never been investigated in Nigeria. This is the maiden research in Nigeria that considers seasonal variation in LSTM neural network model-based $$f_{s}$$ f s forecasting. Accordingly, the novelty and key contribution of our state-of-the-art research is the development and implementation of a low-cost intelligent deep learning model based on the LSTM neural network enhanced with DPA for day-ahead $$f_{s}$$ f s forecasting. To further demonstrate the $$f_{s}$$ f s modelling capability of our technique, we have examined the performances of two different baseline approaches namely, the linear regression (LinReg) model and the adaptive linear element neural network (ADALINE-NN) model. The results of $$f_{s}$$ f s simulation indicated that the proposed LSTM neural network model has the capability to handle varnishing or exploding gradient conundrum. It is highly robust and steady with better accuracy when configured for 24-hour ahead $$f_{s}$$ f s forecasting. The LSTM neural network model outpaced both baseline approaches as model comparisons showed that it has the highest extrapolative accuracy. It presents the lowest RMSE and MAPE and the best NSME and CoC of [2.73, 1.28] m 3 /s, [11.16, 8.16] %, [0.91, 0.83], and [0.97, 0.95] for the rainy and dry season respectively. As the results of the LSTM neural network approach are observed to be more stable in general, it can be established that the proposed model is a practical daily $$f_{s}$$ f s forecasting technique for both the rainy and dry season.

High complexity and randomness in high-speed train-bridge interactive dynamic analysis under earthquake lead to massive calculations in high-speed railway seismic design. To address this challenge, this paper introduces a deep learning-based method at enhancing the computational efficiency of the train-bridge coupled (TBC) system’s seismic response prediction. Accordingly, a deep learning framework for predicting stochastic seismic response is established by using a long short-term memory (LSTM) neural network. Meanwhile, a feasible training strategy for the LSTM network is proposed. A traditional TBC theory and the LSTM neural network are used to perform the seismic response analysis as illustrative examples. The comparative analysis demonstrates the LSTM network's remarkable accuracy and efficiency in predicting the TBC system's seismic response. The prediction performance and extrapolation capability of the LSTM network is evaluated to be good enough to meet the requirement of engineering applications. In contrast to conventional methods like TBC theory or finite element analysis, the LSTM network significantly improves computational efficiency. Furthermore, the computed response data and the evolution characteristics of the probability density function are in good agreement with the conventional method. Therefore, the established LSTM network serves as an effective surrogate model for predicting the TBC system’s seismic response.

Soft sensing of water depth in combined sewers using LSTM neural networks with missing observations

There has been a renewed interest in accurately forecasting the price of crude oil and its fluctuations. That said, this paper aims to study whether the price of crude oil in the United States (US) could be predicted using the stock prices of the top information technology companies. To this end, time-series data was collected and pre-processed as needed, and three architectures of computational neural networks were tested: deep neural networks, long-short term memory (LSTM) neural networks, and a combination of convolutional and LSTM neural networks. The findings suggest that LSTM networks are the best architectures to predict the crude oil price. The outcomes of this paper could potentially help in making the oil price prediction mechanism a more tractable task and in assisting decision-makers to improve macroeconomic policies, generate enhanced macroeconomic projections, and better assess macroeconomic risks.

In order to further explore the application of deep learning in predicting financial market time series data and improve the accuracy of the prediction, this paper adopts a financial time series prediction method based on wavelet denoising, whale optimization algorithm and long-short term memory (LSTM) neural network. This article chooses 10 common evaluation indexes in the financial market as the input, the financial time series data are denoised by wavelet analysis. Then the optimal LSTM neural network parameters are obtained by whale optimization algorithm (WOA). Finally, the LSTM neural network algorithm is used for stock prediction to output the predicted closing price. To verify the effectiveness of WP-WOA-LSTM model, three other neural networks are used to compare with the forecasting result. By comparing the prediction accuracy of different methods, it is obvious that the mean absolute error (MAE) of LSTM neural network under whale optimization algorithm can be reduced by 22 % compared with the standard LSTM neural network. Therefore, the results show that WOA-LSTM model has significantly improved the prediction accuracy.

Artificial neural network is widely used in the financial time series, but Long short-term memory (LSTM) neural network is rarely used in the futures market in China. In this paper, the LSTM neural network is studied by using futures data. The daily trading data of four groups of futures such as silver, copper, lithium and coking coal from December 2014 to December 2018 are used as the training object to make short-term prediction of the closing price. By comparing the Back Propagation (BP) neural network, general multi-layer LSTM neural network, and using the attention mechanism optimization LSTM contrast test, the result of the experiment shows that the futures price trend forecast time sequence, attention mechanism to promote significant effect of time sequence, and LSTM combined effect, by adjusting the parameters setting, using the improved LSTM neural network for time series prediction accuracy is higher, better generalization ability.

As an essential component of urban infrastructure construction, polyethylene (PE) pipelines face the challenging task of underground detection due to the complex and dynamic nature of the subsurface environment, diverse installation paths, and the inherent insulating properties of PE materials. In order to address the non-excavation detection of buried PE pipelines, this paper proposes an acoustic method based on the long short-term memory (LSTM) neural network. The study begins by analyzing the propagation and reflection mechanisms of elastic waves in the pipe-soil coupling system, and a impact excitation source is designed to generate the excitation signal. After establishing the experimental environment and collecting experimental data, a comprehensive analysis is conducted, and the LSTM neural network is employed for data classification to determine the presence of buried PE pipelines. Through neural network training, accurate identification of the PE pipeline’s existence and prediction of its burial depth are achieved, providing an efficient and reliable solution for buried PE pipeline detection. The practical results demonstrate the significant application prospects of the combined acoustic method and LSTM neural network in buried PE pipeline detection. This research contributes a novel solution to the field of non-destructive PE pipeline detection, with both theoretical and practical implications.

Forecasting the short-term metro ridership is an important issue for operation management of metro systems. However, it cannot be solved well by the single long short-term memory (LSTM) neural network alone for the irregular fluctuation caused by various factors. This paper proposes a hybrid algorithm (STL-LSTM) which combines the addition mode of Seasonal-Trend decomposition based on Loess (STL) and the LSTM neural network to mitigate the influences of irregular fluctuation and improve the performance of short-term metro ridership prediction. First, the original series is decomposed into three sub-series by the addition mode of STL. Then, the LSTM neural network is employed to predict each decomposed series. Finally, all the predicted outputs are merged as the overall output. The results show that the STL-LSTM model can achieve higher accuracy than the single LSTM model, support vector regression (SVR), and the EMD-LSTM model which combines the empirical mode decomposition and the LSTM neural network.

The establishment of temperature-displacement correlation is essential for the displacement-based performance evaluation of long-span bridges. We propose a new method based on long short-term memory (LSTM) neural network for modeling multiple temperature-displacement correlation. The LSTM neural network, which can describe the time-lag effect and is suitable for processing ultralong data sequences, is adopted as the basic neural network. We use the adaptive moment estimation method to optimize the LSTM neural network, and introduce the dropout regularization technique to improve the generalization ability of the model. The predominant thermal variables that affect the vertical displacement in the main girder of the bridge are extracted using long-term synchronous monitoring data of temperatures and displacements obtained from a three-span continuous tied arch bridge. An LSTM neural network between multiple temperature variables and displacements is established. A back propagation (BP) neural network is also modeled for comparison. The results show that the structural effective temperature has a significant nonlinear relationship with the vertical displacement in the main girder, while the temperature difference among structural components and the temperature gradient in the main arch rib have a linear correlation with the vertical displacement in the main girder. The effective temperature in the main arch rib and the temperature difference between the main girder and the main arch rib are the predominant thermal variables that result in the vertical displacement in the main girder. Compared with the BP neural network model, the LSTM neural network model proposed in this paper can dramatically reduce the reproduction error and prediction error of the thermal displacement.

In response to the need for an exoskeleton to quickly identify the wearer’s movement mode in the mixed control mode, this paper studies the impact of different feature parameters of the surface electromyography (sEMG) signal on the accuracy of human motion pattern recognition using multilayer perceptrons and long short-term memory (LSTM) neural networks. The sEMG signals are extracted from the seven common human motion patterns in daily life, and the time domain and frequency domain features are extracted to build a feature parameter dataset for training the classifier. Recognition of human lower extremity movement patterns based on multilayer perceptrons and the LSTM neural network were carried out, and the final recognition accuracy rates of different feature parameters and different classifier model parameters were compared in the process of establishing the dataset. The experimental results show that the best accuracy rate of human motion pattern recognition using multilayer perceptrons is 95.53%, and the best accuracy rate of human motion pattern recognition using the LSTM neural network is 96.57%.

In active heave compensation, in order to realize the smooth control of the heave compensation platform, it is necessary to use the ship motion measurement system to accurately obtain the ship displacement signal, invert the ship displacement signal, and then control the expansion and contraction of the electric cylinder so that the compensation platform remains horizontal. The ship displacement measurement system generally adopts the second integral of the acceleration sensor to obtain the ship displacement signal. During the acquisition process of the ship displacement signal, the quadratic integration process of the acceleration, and the communication process of the output control command, there is a processing lag which makes the error accumulate, resulting in a delay in the measurement of the ship motion. In order to collect the ship motion more accurately and control the heave compensation platform more precisely, this paper proposes a ship motion prediction method based on a variable step-variable sampling frequency characteristic LSTM (Long Short-Term Memory) neural network. First, we use the autocorrelation function algorithm to calculate the inherent delay of the lag in the process of signal acquisition by the measurement system. Secondly, the LSTM neural network is used to predict the inherent delay step of the lagging ship displacement signal. During the prediction process, it is found that the difference in the sampling frequency of the displacement signal will lead to a change in the step of the inherent delay—experiment in the laboratory to verify. By controlling the motion platform to simulate the motion of the ship and using the ship motion measurement system and the laser sensor system to measure the displacement signal of the motion platform synchronously, it is verified that the ship motion measurement system does have an inherent delay. Thirdly, on a sailing ship, ship displacement signals are collected by setting multiple sets of ship motion measurement systems. Finally, multiple sets of sampling frequency and multiple steps are set, and the ship motion is predicted based on the variable step-variable sampling frequency LSTM neural network. It is verified that the prediction accuracy is related to the sampling frequency of the signal collector and the prediction step of the LSTM neural network, which improves the prediction accuracy of the model and the timeliness of ship motion acquisition.

Forecasting building energy consumption: Adaptive long-short term memory neural networks driven by genetic algorithm

In the face of the rapid growth and development of regional integrated energy system (RIES) globally, accurate load prediction technique is increasingly playing a critical role in RIES planning. This paper presents a Bayesian Optimized Long Short-Term Memory (BO-LSTM) neural network to predict the electric, heating and cooling power load for the short and mid-term operation. The Bayesian optimization algorithm is performed to automate hyperparameter tuning to improve results, so avoiding different hyperparameters may lead to considerable differences in the performance of other deep learning network architecture in some sense. The developed model is validated on one actual RIES in China for data collected in a year. The simulation results of the proposed BO-LSTM indicate the effectiveness and excellent prediction accuracy in comparison with other traditional models, such as autoregressive integrated moving average model (ARIMA), long short-term memory (LSTM) and convolutional neural network (CNN).

Since it is one of the world's most significant financial markets, the foreign exchange (Forex) market has attracted a large number of investors. Accurately anticipating the forex trend has remained a popular but difficult issue to aid Forex traders' trading decisions. It is always a question of how precise a Forex prediction can be because of the market's tremendous complexity. The fast advancement of machine learning in recent decades has allowed artificial neural networks to be effectively adapted to several areas, including the Forex market. As a result, a slew of research articles aimed at improving the accuracy of currency forecasting has been released. The Long Short-Term Memory (LSTM) neural network, which is a special kind of artificial neural network developed exclusively for time series data analysis, is frequently used. Due to its high learning capacity, the LSTM neural network is increasingly being utilized to predict advanced Forex trading based on previous data. This model, on the other hand, can be improved by stacking it. The goal of this study is to choose a dataset using the Hurst exponent, then use a two-layer stacked Long Short-Term Memory (TLS-LSTM) neural network to forecast the trend and conduct a correlation analysis. The Hurst exponent (h) was used to determine the predictability of the Australian Dollar and United States Dollar (AUD/USD) dataset. TLS-LSTM algorithm is presented to improve the accuracy of Forex trend prediction of Australian Dollar and United States Dollar (AUD/USD). A correlation study was performed between the AUD/USD, the Euro and the Australian Dollar (EUR/AUD), and the Australian Dollar and the Japanese Yen (AUD/JPY) to see how AUD/USD movement affects EUR/AUD and AUD/JPY. The model was compared with Single-Layer Long Short-Term (SL-LSTM), Multilayer Perceptron (MLP), and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise–Improved Firefly Algorithm Long Short-Term Memory. Based on the evaluation metrics Mean Square Error (MSE), Root Mean Square Error, and Mean Absolute Error, the suggested TLS-LSTM, whose data selection is based on the Hurst exponent (h) value of 0.6026, outperforms SL-LSTM, MLP, and CEEMDAN-IFALSTM. The correlation analysis conducted shows both positive and negative relations between AUD/USD, EUR/AUD, and AUD/JPY which means that a change in AUD/USD will affect EUR/AUD and AUD/JPY as recorded depending on the magnitude of the correlation coefficient (r).