AUTOMATED INFORMATION SYSTEM FOR TRAINING RESCUERS AND INDUSTRIAL PERSONNEL TO ACT IN EXTREME SITUATIONS

An automated information system for assessing the physiological readiness of specialists to perform tasks in conditions of limited breathing mixture resources is presented. The purpose of the system is to increase the reliability and safety of personnel actions in extreme situations through personalized prediction of physiological reactions. The system implements a hybrid model of intelligent data analysis that predicts the time series of pulmonary ventilation based on route characteristics, physical exertion, and individual user characteristics. The model is based on long short-term memory sequence-to-sequence architecture and includes a trainable spectral-inertial output signal smoothing layer. A modified loss function is proposed that takes into account error asymmetry and data timeliness, which improves the accuracy and reliability of the forecast. The system is implemented as a web application with an interface for setting routes, selecting users and visualizing results. Experimental tests have confirmed the high accuracy of the model and its resistance to individual variability, demonstrating the effectiveness of the proposed approach for training and assessing staff readiness.