جستجوی مقالات مرتبط با کلیدواژه « Long Short-Term Memory (LSTM) » در نشریات گروه « مکانیک »

This paper presents a real-time approach for detecting compensatory movements in upper limb rehabilitation for stroke patients using deep learning algorithms. The study applied Recurrent Neural Networks (RNN), Gated Recurrent Unit (GRU), Long-Short-Term-Memory (LSTM), and Transformer to analyze Microsoft Kinect data from the Toronto Rehab Stroke Pose dataset. The models were trained with focal loss to address imbalanced data distribution. The simulation results showed that the proposed deep learning algorithms are effective in detecting compensatory movements. The GRU-based models provide the fastest results and the transformer models exhibit the best accuracy and fastest inference time on the employed CPU .

This paper presents a real-time approach for detecting compensatory movements in upper limb rehabilitation for stroke patients using deep learning algorithms. The study applied Recurrent Neural Networks (RNN), Gated Recurrent Unit (GRU), Long-Short-Term-Memory (LSTM), and Transformer to analyze Microsoft Kinect data from the Toronto Rehab Stroke Pose dataset. The models were trained with focal loss to address imbalanced data distribution. The simulation results showed that the proposed deep learning algorithms are effective in detecting compensatory movements. The GRU-based models provide the fastest results and the transformer models exhibit the best accuracy and fastest inference time on the employed CPU .

Renewable sources for power generation are being popular day by day and solar PV is mostly first choice. In addition, the energy output of solar PV is highly affected by weather conditions like temperature, irradiance, sky conditions etc. Therefore, an intelligent model based on weather conditions is essential for estimation of solar energy output to meet the needs of energy required. The prediction of PV power output is critical to security, operation, scheduling and energy management. Stability of power grid can also be increased if accuracy of power production in PV plants is further enhanced. This paper has worked on LSTM and used recurrent neural networks (RNN) for forecasting of power production and it is seen that the results of RNNs are nearly compatible with the realistic power production which is evident from less mean absolute error (MAE), mean absolute percentage error (MAPE), Root Mean Square Percentage Error (RMPSE) of magnitude. The comparison with different layers of LSTM model for each season of weather is analyzed

Renewable sources for power generation are being popular day by day and solar PV is mostly first choice. In addition, the energy output of solar PV is highly affected by weather conditions like temperature, irradiance, sky conditions etc. Therefore, an intelligent model based on weather conditions is essential for estimation of solar energy output to meet the needs of energy required. The prediction of PV power output is critical to security, operation, scheduling and energy management. Stability of power grid can also be increased if accuracy of power production in PV plants is further enhanced. This paper has worked on LSTM and used recurrent neural networks (RNN) for forecasting of power production and it is seen that the results of RNNs are nearly compatible with the realistic power production which is evident from less mean absolute error (MAE), mean absolute percentage error (MAPE), Root Mean Square Percentage Error (RMPSE) of magnitude. The comparison with different layers of LSTM model for each season of weather is analyzed.