Predictive Diagnostic and State Classification of the Inertial Navigation System Using ML

The problem of inertial navigation system (INS) state classification and remaining useful life prediction is considered within the framework of the Machine learning approach; the proposed LSTM and gradient boosting algorithms for its solution is described. The initial data is divided into two types: normal and abnormal. Both types of data contain information about temperature readings measured by INS sensors. The analysis uses both absolute temperature values and its changes over time, which makes it possible to identify long-term trends and short-term fluctuations associated with the operational features of the device. So the first task is formulated as a binary classification task. If the INS state is classified as abnormal the RUL prediction task is solved. The advantages of the proposed algorithms are discussed, and their results received on INS real temperature data are analyzed. So the main goal of the paper is to develop and test a classification model that will make it possible to quickly and accurately determine the INS state based on the analysis of current and predicted temperature values, thereby helping to increase the reliability and efficiency of the predictive maintenance system.