A Study on Disturbance Classification of Unmanned Vehicle Data Using SVM

In this paper, we present disturbance classification from unmanned vehicle data using a support vector machine (SVM). Disturbances that occur while operating unmanned vehicles are classified into various categories. We considered a problem in which the unmanned vehicle was a quadrotor model, and the type of operation data that could be used to classify disturbances well was selected by considering the characteristics of the model. Time series data of unmanned vehicle operation are not suitable for solving using SVM, necessitating the extraction of statistical features from time series data that can be well classified. Accordingly, we modeled a quadrotor and various disturbances to investigate unmanned vehicle disturbance classification using an SVM with selected statistical features. Experiments show that the statistical features and SVM can classify the disturbances well. © ICROS 2022.