Predicting UAV Type: An Exploration of Sampling and Data Augmentation for Time Series Classification
This work addresses safety concerns in restricted airspace by enabling UAV type prediction, but it is incremental as it applies standard methods (LSTM with sampling/augmentation) to a specific domain.
The paper tackled the problem of classifying unmanned aerial vehicle types (quadrotor, hexarotor, fixed-wing) using an LSTM neural network for time series classification, achieving macro f-scores of 98.16% for quadrotor, 73.15% for fixed-wing, and 42.15% for hexarotor despite class imbalance.
Unmanned aerial vehicles are becoming common and have many productive uses. However, their increased prevalence raises safety concerns -- how can we protect restricted airspace? Knowing the type of unmanned aerial vehicle can go a long way in determining any potential risks it carries. For instance, fixed-wing craft can carry more weight over longer distances, thus potentially posing a more significant threat. This paper presents a machine learning model for classifying unmanned aerial vehicles as quadrotor, hexarotor, or fixed-wing. Our approach effectively applies a Long-Short Term Memory (LSTM) neural network for the purpose of time series classification. We performed experiments to test the effects of changing the timestamp sampling method and addressing the imbalance in the class distribution. Through these experiments, we identified the top-performing sampling and class imbalance fixing methods. Averaging the macro f-scores across 10 folds of data, we found that the majority quadrotor class was predicted well (98.16%), and, despite an extreme class imbalance, the model could also predicted a majority of fixed-wing flights correctly (73.15%). Hexarotor instances were often misclassified as quadrotors due to the similarity of multirotors in general (42.15%). However, results remained relatively stable across certain methods, which prompted us to analyze and report on their tradeoffs. The supplemental material for this paper, including the code and data for running all the experiments and generating the results tables, is available at https://osf.io/mnsgk/.