Machine Learning Boosting Algorithms for Determining Euler Angles in an Inertial Navigation System

Many countries of the world are actively researching and developing a large number of marine equipment for the study and development of the oceans. Along with the classical methods of ocean exploration, autonomous platforms are actively developed, such as autonomous underwater vehicles, gliders and autonomous surface vehicles. The article is devoted to improving the accuracy of the inertial navigation system of such platforms based on MEMS sensors using machine learning methods. We used two boosting methods for this: adaptive boosting and gradient boosting. Accuracy obtained by gradient boosting algorithm exceeds the level of adaptive boosting: $0.913 \pm 0.006$ on the training set and 0.912 on the test set. Adaptive boosting shows its best performance with more individual trees, which leads to more time to train the model and calculate the result.