Implementasi Long Short Term Memory (LSTM) dalam Deteksi Kantuk pada Pengemudi Menggunakan Sensor Detak Jantung

Traffic accidents are often caused by drowsiness or negligent sleep, as well the use of alcohol or drugs. Microsleep, which is drowsiness or falling asleep within a few seconds without the driver realizing it, is a dangerous condition that can lead to death while driving. This research aims to implement the Long Short Term Memory (LSTM) algorithm as an early warning of microlsleep in drivers and develop a drowsiness detection tool using a pulse heart rate sensor. LSTM, with its ability to memory long-range information, has proven to be superior in time series prediction and is applied in real-time driver heart rate data analysis. The results show that the implemented LSTM model has good performance in detecting drowsiness, with MAE values of 6.42 in training data and 6.35 in testing data. RMSE of 8.82 for training and 8.33 for testing. MAPE of 8.87% in training data and 8.97% in testing data, and MSE of 77.80 in training data and 69.47 in testing. Thus, the LSTM algorithm is effective in detecting drowsiness in drivers through heart rate data analysis, which can serve as an early warning system to prevent traffic accidents caused by microsleep.