Memory Deep Neural Network Research Articles

LPM-Net: A Data-Driven Resource-Efficient Predictive Motion Planner for Mobile Robots

A data-driven predictive motion planner for mobile robots, referred to as LPM-Net, has been proposed in this paper. Conventional predictive motion planners are computationally expensive, often resulting in insufficient throughput on mobile robot hardware. LPM-Net is an imitation learning-assisted local predictive non-holonomic motion planner that is capable of learning from conventional motion planners regarded as paradigm models and replicating their behavior while satisfying the same kinodynamic constraints. In addition, LPM-Net is compatible with GPU and TPU hardware, allowing for faster and more efficient processing. LPM-Net uses convolutional and recurrent long short-term memory deep neural networks to predict steering commands. This has improved computational efficiency which allows autonomous vehicles to be equipped with more cost-effective computers. In the present study, LPM-Net was tuned to mimic the behavior of a model predictive controller paradigm model. Measurements in this study demonstrate that the proposed mimic planner, LPM-Net, consumes approximately half the processing power of the conventional predictive planner, albeit with a slight increase in hesitation when reaching goals.

Deep Learning Model for Prediction of Dementia Severity based on EEG Signals

This study aimed to determine variations in the electroencephalograms (EEGs) of 15 individuals who were diagnosed with mild cognitive impairment (MCI) following stroke, 5 individuals suffering from vascular dementia (VD) and 15 healthy normal control (NC) individuals who performed a working memory task. Conventional filters including notch and bandpass filters were utilised to remove noise from the EEG data. The proposed method comprises computing the estimates of the attention entropy (AttEn), bubble entropy (BubbEn) and symbolic dynamic entropy (SyDyEn) of univariate data sequence features. The long short-term memory (LSTM) deep learning neural network was used to automatically classify dementia severity through noninvasive EEG-based recordings. The LSTM classification result with AttEn exceeds an average of 88.9% than BubbEn and SyDyEn, with classification results of 69.2% and 77.7%, respectively. The analysis of the brain EEG-based dementia severity dataset suggests that AttEn could potentially serve as a biomarker for detecting dementia severity. AttEn can capture relevant patterns and features in the EEG data and may be indicative of the severity of dementia with LSTM RNN to differentiate patients with VD, patients with MCI and NC individuals.

Harnessing Deep Learning for Enhanced MPPT in Solar PV Systems: An LSTM Approach Using Real-World Data

Maximum Power Point Tracking (MPPT) is essential for maximizing the efficiency of solar photovoltaic (PV) systems. While numerous MPPT methods exist, practical implementations often lean towards conventional techniques due to their simplicity. However, these traditional methods can struggle with rapid fluctuations in solar irradiance and temperature. This paper introduces a novel deep learning-based MPPT algorithm that leverages a Long Short-Term Memory (LSTM) deep neural network (DNN) to effectively track maximum power from solar PV panels, utilizing real-world data. The simulations of three algorithms—Perturb and Observe (P&O), Artificial Neural Network (ANN), and the proposed LSTM-based MPPT—were conducted using MATLAB (2021b) and RT_LAB (24.3.3) with an OPAL-RT simulator for real-time analysis. The data used for this study were sourced from NASA/POWER’s Native Resolution Daily Data of solar irradiation and temperature specific to Imphal, Manipur, India. The obtained results demonstrate that the LSTM-based MPPT system achieves a superior power tracking accuracy under changing solar conditions, producing an average output of 74 W. In comparison, the ANN and P&O methods yield average outputs of 57 W and 62 W, respectively. This significant improvement, i.e., 20–30%, underscores the effectiveness of the LSTM technique in enhancing the power output of solar PV systems. By incorporating real-world data, valuable insights into solar power generation specific to the selected location are provided. Furthermore, the outputs of the model were verified through real-time simulations using the OPAL-RT simulator OP4510, showcasing the practical applicability of this approach in real-world scenarios.

Application Research of Power Measurement Data Exchange Based on Uninterruptible Power Supply

With the maturity of the smart meter embedded system integration mode, the power consumption information can be fed back to the management center by the intelligent terminal through the power Internet of Things. The smart terminal adopts the narrowband data communication mode. However, the narrowband channel interval is hindered by interference signals, resulting in serious data packet loss. Therefore, a stable and efficient acquisition scheme is needed to ensure its smooth interaction. Based on the application of smart meter data communication, we analyze the power consumption optimization and carrier transmission optimization of the power Internet of Things and propose a data analysis method based on self-learning and external characteristics, which is used to calculate the offline execution strategy of smart meters when they encounter factors such as power outage or intrusion. A production task-driving rule is designed. The external characteristics of deep learning data of the power Internet of Things are collected. The accuracy of data classification is judged according to the parallel Internet of Things of a convolutional neural network and a long-term and short-term memory deep neural network. When dealing with high-dimensional original data, the data network is graded to reduce the work of feature selection and the difficulty of training. Moreover, we improve the autonomous execution ability of an embedded terminal device in an offline state. We also strengthen the ability of abnormal data detection and judgment learning to meet the requirements of efficient and stable operation of the power grid. Simulation results verify that compared with the current mainstream schemes, the performance of the proposed scheme is improved by more than 17%. Besides, the fault tolerance rate is enhanced by 29%, and the power consumption after optimization is reduced by 26.3% and 42.8%. These effects highlight the role of data processing in embedded systems. It solves the problem of abnormal data and energy loss affecting the accuracy of error estimation of electric energy meters, opens up a new way for data acquisition and processing of large-scale smart meters, and further improves the service level of information management of power systems.

Real-time Blood Pressure Prediction on Wearable Devices using Edge Based Deep Neural Networks: A Hardware-software Co-design Approach

This paper presents the hardware realization of a real-time blood pressure (BP) prediction model for wearable devices, utilizing long short-term memory (LSTM) deep neural networks (DNNs). The proposed system uses both electrocardiogram (ECG) and photoplethysmogram (PPG) signal values for BP prediction. It aims to address the limitations of traditional BP measurement methods, providing a low error, minimal computational overhead, more accurate and convenient alternative system for individuals with hypertension or at risk for cardiovascular diseases. The utilization of split matrix approach leads to a reduction in hardware complexity across the entire system. This technique involves breaking down the larger weight matrices used in the computations of DNNs into smaller matrices. This fragmentation results in a decrease in the complexity of the hardware responsible for performing matrix vector multiplications (MVMs) within LSTMs. The resultant architecture of the predictive model gains several advantages, including a lowered level of complexity in terms of the space occupied by individual cells, decreased processing delay, and reduced power consumption. Furthermore, this approach enables the achievement of a notably improved minimum achievable clock period of 2.972 ns. This prediction model can operate locally on wearable devices, reducing the reliance on cloud computing and improving privacy and security. The performance evaluations are carried out using both analytical and implementation results. The results indicate that the proposed model can be practically applied to real-world problems and can potentially enhance the accuracy of various machine-learning tasks.

Research on numerical modeling of two-dimensional freak waves and prediction of freak wave heights based on LSTM deep learning networks

Freak waves have posed a serious threat to the safety of marine structures. Thus, the accurate simulation and prediction of freak waves are crucial for maritime safety. This paper proposes a nonlinear numerical method based on phase modulation. This method achieves precise time and location simulation of two-dimensional freak waves, while preserving the key statistical characteristics of the wave sequence, thereby obtaining more accurate two-dimensional wave height information. Additionally, this paper constructs a Long Short-Term Memory (LSTM) deep neural network integrated with a Sequence-to-Sequence (seq2seq) framework for predicting sequences of freak wave heights. In this study, we conduct an in-depth predictive analysis of the experimental data for sequences of freak wave heights. And this paper provides a comprehensive assessment of the performance of Recurrent Neural Networks (RNNs), Gated Recurrent Units (GRUs), and Long Short-Term Memory (LSTM) networks in the task of freak wave prediction. The experimental results indicate that the seq2seq-LSTM model demonstrates significant performance advantages in predicting the wave heights of freak waves, particularly for long-term predictions. In summary, the numerical simulation method based on phase modulation and the improved seq2seq-LSTM deep learning model are of significant practical value for enhancing the safety of offshore platforms and ships.

Deep learning approach for predicting monsoon dynamics of regional climate zones of India

The complex interplay of various complicated meteorological and oceanic processes has made it more difficult to accurately predict Indian monsoon rainfall. A future-oriented and one of the most potential methods for predictive analytics is deep learning. The proposed work exploits empirical Mode Decomposition-Detrended Fluctuation Analysis (EMD-DFA) and long short-term memory (LSTM) deep neural networks (EMD-LSTM) to build novel predictive models and analyze predictability effectively. The time series data of each homogeneous monsoon zone are decomposed into different empirical time series components known as intrinsic mode functions (IMFs). The proposed work's obtained results report that the EMD-LSTM hybrid strategy consistently outperforms other methods in terms of accuracy. Furthermore, we examined possible relationships between each homogeneous monsoon zone and multiple climate drivers, shedding light on the complicated relationships that influence monsoon patterns. This study presents a unique way of predicting complex monsoon rainfall in homogenous regions of India and marks the first application of the EMD-LSTM technique for this purpose to the best of our knowledge which is necessary for improving water conservation and distribution at different climate zones of India.

Modified receiver architecture in software-defined radio for real-time modulation classification

Automatic modulation classification (AMC) is an important process for future communication systems with prominent applications from spectrum management, and secure communication, to cognitive radio. The requirement for an efficient AMC classifier is due to its capability in blind modulation recognition, which is a difficult task in real scenarios where the limitations of traditional hardware and the complexity of channel impairments are involved. Therefore, this paper proposes a complete real-time AMC system based on software-defined radio and deep learning architecture. The system demodulation performance is verified through simulations and real channel impairment conditions to ensure reliability. With at most 6 times reduced number of parameters, two proposed models convolutional long short-term memory deep neural network and residual long short-term memory neural network also show a general improvement in classification accuracy compared with reference studies. The performance of these models at real-time AMC is tested with suitable processing time for practical applications.

CNN-BiLSTM-DNN-Based Modulation Recognition Algorithm at Low SNR

Radio spectrum resources are very limited and have become increasingly tight in recent years, and the exponential growth of various frequency-using devices has led to an increasingly complex and changeable electromagnetic environment. Wireless channel complexity and uncertainty have increased dramatically, and automated modulation recognition (AMR) performs poorly at low signal-to-noise ratios. It is proposed to use convolutional bidirectional long short-term memory deep neural networks (CNN-BiLSTM-DNNs) as a deep learning framework to extract features from single and mixed in-phase/orthogonal (I/Q) symbols in modulated data. The framework combines the capabilities of one- and two-dimensional convolution, a bidirectional long short-term memory network, and a deep neural network more efficiently, extracting characteristics from the perspective of time and space to enhance the accuracy of automatic modulation recognition. Modulation recognition experiments on the benchmark datasets RML2016.10b and RML2016.10a show that the average recognition accuracies of the proposed model from −20 dB to 18 dB are 64.76% and 62.73%, respectively, and the improvement ranges of modulation recognition accuracy are 0.29−5.56% and 0.32−4.23% when the signal-to-noise ratio (SNR) is −10 dB to 4 dB, respectively. The CNN-BiLSTM-DNN model outperforms classical models such as MCLDNN, MCNet, CGDNet, ResNet, and IC-AMCNet in terms of modulation type recognition accuracy.

Long short-term memory (LSTM) neural networks for predicting dynamic responses and application in piezoelectric energy harvesting

Long Short-Time Memory (LSTM) deep neural networks are capable of learning order dependence in sequence problems and capturing long-term, non-linear temporal dependencies between the input and out of a system. With the long-term vision to model dynamical systems to which analytical or numerical methods are impossible or difficult to apply, this paper presents a study of modeling system dynamics and predicting responses using the LSTM networks, which have demonstrated excellent capability in predicting single-mode responses in a prior study. However, the LSTM network exhibits difficulties in modeling and predicting multi-mode responses accurately. To resolve the multi-mode issue, this paper presents an approach that obtains an equivalent network consisting of a set of sub-networks learned on isolated modes, and demonstrates its effectiveness on a simulated 2-degree-of-freedom mass-spring-damper system of nonlinear Duffing springs. The second part of the paper is focused on the application of the proposed approach in piezoelectric energy harvesting. Experiments are conducted on a harvester subjected to random base-motion excitation and exhibiting nonlinearity in its multi-mode response. Both the direct and mode-separation LSTM modeling approaches are applied to predict the output voltage given a random base-motion excitation. The mode-separation approach outperforms the direct approach significantly, and yields an excellent match between the actual and predicted responses. Specifically, for a test electrical voltage response of RMS value 0.2241 V, the difference between the actual test and predicted responses by using the mode-separation approach has an RMS value of 0.0504 V, compared to 0.1645 V obtained by using the direct LSTM approach. It is also much lower than the RMS value of 0.1835 V obtained by using the attention-based LSTM network, another comparison method. Leveraging a deep learning strategy, the validated approach opens up opportunities for accurately modeling energy harvesting systems of high complexities and/or strong nonlinearities.

Advancing Accuracy in Sea Level Estimation with GNSS-R: A Fusion of LSTM-DNN-Based Deep Learning and SNR Residual Sequences

The global navigation satellite system reflectometry (GNSS-R) technique has shown promise in retrieving sea levels using signal-to-noise ratio (SNR) data. However, its accuracy and performance are often limited compared to conventional tide gauges, particularly due to constraints in satellite elevation angles. To address these limitations, we propose a methodology integrating Long Short-Term Memory Deep Neural Networks (LSTM-DNN) models, utilising SNR residual sequences as key feature inputs. Our study focuses on the SC02 station, examining elevation angles ranging from 5° to 10°, 5° to 15°, and 5° to 20°. Results reveal notable reductions in root mean square errors (RMSE) of 2.855%, 17.519%, and 15.756%, respectively, showcasing improvements in accuracy across varying elevation angles. Of particular significance is the enhancement in precision observed at higher elevation angles. This underscores the valuable contribution of our approach to nearshore sea level wave height retrieval, promising advancements in the GNSS-R technique.

Forecasting cryptocurrency returns using classical statistical and deep learning techniques

The emergence of cryptocurrencies has generated enthusiasm and concern in the modern global economy. However, their high volatility, erratic price fluctuations, and tendency to exhibit price bubbles have made investors cautious about investing in them. Consequently, it is essential to develop methods and models to forecast cryptocurrency returns to benefit investors, traders, and the scientific community. Despite the considerable volume of research on Bitcoin price forecasting, other cryptocurrencies have received little attention in academic literature. Additionally, the current body of literature on predicting cryptocurrency prices or returns emphasizes the use of in-sample methodologies. However, this method is susceptible to overfitting. To address these gaps in the literature, this study employs autoregressive moving average (ARMA), generalized autoregressive conditional heteroskedasticity (GARCH), exponential generalized autoregressive conditional heteroskedasticity (EGARCH), and long short-term memory (LSTM) deep learning neural networks to forecast returns for the ten most actively traded digital currencies: Bitcoin, Ethereum, Ripple, Chainlink, Litecoin, Cardano, Ethereum Classic, Bitcoin Cash, Tether, and Binance Coin. To assess the accuracy of the two models, this study utilizes an out-of-sample method with data gathered sequentially from November 9, 2017, to September 18, 2022. The results indicate that all models exhibit high accuracy, as evidenced by their low root mean square error (RMSE), mean absolute error (MAE), and mean squared error (MSE) values. Meanwhile, the hybrid EGARCH-LSTM or GARCH-LSTM models demonstrate slightly better accuracy compared with the other models. The findings are valuable for investors, traders, and researchers involved in cryptocurrency forecasting.

Analyse Chinese Teaching Project-Based Learning Design Application Optimization Using Keras Model

Educational methods for data mining may be used to successfully analyze data from a variety of educational platforms, including e-learning, e-admission programs, and automated results management solutions, to provide very helpful insights into students' performance. Chinese language teaching will reach new heights as a result of the rapid advancement of network technology and the constant improvement of network transmission speed. Additionally, the acceptance of online payment methods is growing, simplifying international and web-based trade. Deep learning models have become more significant in many spheres of life recently. It made it possible for investigators to automatically extract superior characteristics from unprocessed data.With the aid of the Keras model, make it is easier for students to rapidly obtain practical knowledge with deep learning structures and boost their confidence and enthusiasm for learning. In this study, we investigate the Long Short-Term Memory (LSTM) deep neural network model with Novel Swarm Optimization (NSO) network to accurately forecast student improvement using historical data. The most cutting-edge LSTM and an optimization mechanism model (NSO) have been employed in this work to solve research issues based on sophisticated feature categorization and prediction. For Chinese academics, institutions, and government agencies to accurately anticipate performance, this study is very important. When paired with the optimization mechanism (NSO), LSTM's better sequence learning capabilities provide performance that is superior to the current state-of-the-art. The prediction accuracy of the suggested approach is 93%.

Research on Communication Signal Modulation Recognition Based on a CCLDNN

In this paper, a new automatic modulation recognition (AMR) method named CCLDNN (complex-valued convolution long short-term memory deep neural network) is proposed. It is designed to significantly improve the recognition accuracy of modulation modes in low signal-to-noise ratio (SNR) environments. The model integrates the advantages of existing mainstream neural networks. The phase and amplitude information of complex signals is effectively captured through a complex module in the input layer. The Squeeze-and-Excitation (SE) attention mechanism, Bi-LSTM layer, and deep convolutional layer are introduced in the feature extraction layer to gradually enhance feature expression. Among these, the introduction of LSTM enables the model to capture the sequence dependence of signals, and the application of the SE attention mechanism further improves the model’s ability to focus on key features. Tests using the RadioML2016.10a dataset show that the model performs well at multiple SNR levels, achieving an average recognition accuracy of more than 80% over an SNR range of 0 dB to 18 dB. However, under the condition of a low SNR from −20 dB to −2 dB, the model still maintains a high recognition ability. The advanced CCLDNN method shows great deep learning potential in solving practical communication problems.

Hybrid Neural Networks for Enhanced Predictions of Remaining Useful Life in Lithium-Ion Batteries

With the proliferation of electric vehicles (EVs) and the consequential increase in EV battery circulation, the need for accurate assessments of battery health and remaining useful life (RUL) is paramount, driven by environmentally friendly and sustainable goals. This study addresses this pressing concern by employing data-driven methods, specifically harnessing deep learning techniques to enhance RUL estimation for lithium-ion batteries (LIB). Leveraging the Toyota Research Institute Dataset, consisting of 124 lithium-ion batteries cycled to failure and encompassing key metrics such as capacity, temperature, resistance, and discharge time, our analysis substantially improves RUL prediction accuracy. Notably, the convolutional long short-term memory deep neural network (CLDNN) model and the transformer LSTM (temporal transformer) model have emerged as standout remaining useful life (RUL) predictors. The CLDNN model, in particular, achieved a remarkable mean absolute error (MAE) of 84.012 and a mean absolute percentage error (MAPE) of 25.676. Similarly, the temporal transformer model exhibited a notable performance, with an MAE of 85.134 and a MAPE of 28.7932. These impressive results were achieved by applying Bayesian hyperparameter optimization, further enhancing the accuracy of predictive methods. These models were bench-marked against existing approaches, demonstrating superior results with an improvement in MAPE ranging from 4.01% to 7.12%.

A Softwarized Intrusion Detection System for IoT-Enabled Smart Healthcare System

The Internet of Things-enabled Smart Healthcare System (IoT-SHS) is a networked infrastructure of intelligent wearables, software applications, health systems, and services that continuously monitors and transmits patient-sensitive data using an open wireless channel. The conventional security mechanisms are unsuitable for detecting attacks in the dynamic IoT-SHS context due to resource limitations and heterogeneity in low-cost healthcare devices. Deep Learning (DL) solutions for Intrusion Detection System (IDS) and softwarization of the network has the potential to achieve secure network services in the IoT-SHS environment. Motivated by the aforementioned discussion, we propose an intelligent softwarized IDS for protecting the critical infrastructure of the IoT-SHS ecosystem. Specifically, the DL-based IDS is designed using a hybrid cuda Long Short-Term Memory Deep Neural Network (cuLSTM-DNN) algorithm to assist network administrators in efficient decision-making for the generated intrusions. To further bolster the system’s resilience, we suggest a deployment architecture for the proposed CUDA-powered IDS using OpenStack Tacker in a real SDN environment, ensuring that virtual machines can directly utilize the host’s NVIDIA GPU, thereby streamlining and enhancing the network’s operational efficiency. The experimental results using the CICDDoS2019 dataset confirm the effectiveness of the proposed framework over some baseline and recent state-of-the-art techniques.

Fault Diagnosis of the LAMOST Fiber Positioner Based on a Long Short-term Memory (LSTM) Deep Neural Network

The Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) has been in normal operation for more than 10 yr, and routine maintenance is performed on the fiber positioner every summer. The positioning accuracy of the fiber positioner directly affects the observation performance of LAMOST, and incorrect fiber positioner positioning accuracy will not only increase the interference probability of adjacent fiber positioners but also reduces the observation efficiency of LAMOST. At present, during the manual maintenance process of the positioner, the fault cause of the positioner is determined and analyzed when the positioning accuracy does not meet the preset requirements. This causes maintenance to take a long time, and the efficiency is low. To quickly locate the fault cause of the positioner, the repeated positioning accuracy and open-loop calibration curve data of each positioner are obtained in this paper through the photographic measurement method. Based on a systematic analysis of the operational characteristics of the faulty positioner, the fault causes are classified. After training a deep learning model based on long short-term memory, the positioner fault causes can be quickly located to effectively improve the efficiency of positioner fault cause analysis. The relevant data can also provide valuable information for annual routine maintenance methods and positioner designs in the future. The method of using a deep learning model to analyze positioner operation failures introduced in this paper is also of general significance for the maintenance and design optimization of fiber positioners using a similar double-turn gear transmission system.

Improved fault diagnosis method of electric gate valve in nuclear power plant

The vibration signal (VS) analysis and acoustic emission signal (AES) analysis methods are the common effective diagnostic methods in valve fault diagnosis (FD)(Classify fault types). However, these two methods are disadvantaged by the surrounding environment and noise signals produced during acquisition. To obtain an improved FD effect of an electric gate valve (EGV) in nuclear power plant, the two previous methods are ameliorated by a noise reduction processing, then the signal feature extraction is performed. The FD and degree evaluation are carried out for the common fault types of the EGV taken as object. In this paper, three-phase unbalanced faults (Acquisition of VS), packing damage faults (Acquisition of VS) and internal leakage faults (Acquisition of AES) are taken as examples. The Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) algorithm composed with Fast Fourier Transform (FFT) and the SAEU3H series digital acoustic emission detection system (SAEU3H SDAEDS) are used to noise reduction and extraction of parameters of the VS and AES, respectively. Then, the extracted feature parameters are brought into the Bi-directional Long Short-Term Memory (Bi-LSTM) deep neural network for FD and fault degree evaluation (Assess the severity of failure occurrences). The experimental results show that the method adopted in this paper has high FD accuracy and low fault evaluation error.

Influence-Based Nano Fuzzy Swarm Oxygen Deficiency Detection and Therapy

Oxygen deficiency is a serious health problem that may occur as a result of many diseases. In this article, we present an influence-based nano fuzzy swarm (INFS) for oxygen deficiency detection and therapy using a swarm of oxygen carrier nanomachines operating in three cognitive fields of control, influence, and interest. In particular, we propose a long short-term memory (LSTM) deep neural network for detecting apnea by analyzing the irregular peripheral oxygen saturation (SpO2) signal. Using the proposed sleep-in-the-loop strategy and the desaturated blood biomarkers, including oxygen and hydrogen ion concentrations, an in-silico multithreshold nano fuzzy swarm noninvasive therapeutic method is then performed. We also analytically prove the stability of the INFS using swarm control theory. We apply our strategy to sleep apnea, as one of the most common instances of oxygen deficiency. Furthermore, we compare the accuracy of INSF by using LSTM, bidirectional LSTM (BiLSTM), multilayer perceptron (MLP), convolutional neural network (CNN), and support vector machines (SVM). The detection and therapy results are then compared with other apnea detection methods. The input variables and structure of INSF, i.e., the number of rules and width of membership functions, are studied in terms of robustness to noise. As the results show, the proposed artificial intelligence (AI)-based noninvasive nano detection and therapy method could outperform the competing approaches in treating oxygen deficiency emergencies such as apnea.

Fatigue Assessment from Facial Videos using Deep Neural Networks and Engineered Features Informed by Domain Knowledge.

Fatigue impairs cognitive and motor function, potentially leading to mishaps in high-pressure occupations such as aviation and emergency medical services. The current approach is primarily based on self-assessment, which is subjective and error-prone. An objective method is needed to detect severe and likely dangerous levels of fatigue quickly and accurately. Here, we present a quantitative evaluation tool that uses less than two minutes of facial video, captured using an iPad, to assess fatigue vs. alertness. The tool is fast, easy to use, and scalable since it uses cameras readily available on consumer-electronic devices. We compared the classification performance between a Long Short-Term Memory (LSTM) deep neural network and a Random Forest (RF) classifier applied to engineered features informed by domain knowledge. The preliminary results on an 11-subject dataset show that RF outperforms LSTM, with added interpretability on the features used. For the RF classifiers, the average areas under the receiver operating characteristic curve, based on the 11-fold and individualized 11-fold cross validations, are 0.72 ± 0.16 and 0.8 ± 0.12, respectively. Equal error rates are 0.34 and 0.26, respectively. This study presents a promising approach for rapid fatigue detection. Additional data will be collected to assess the generalizability across populations.