(An) artificial neural network-based model for supporting UAV path prediction and autonomous flight control

Abstract

Recently, UAVs have been commercialized for military companies and civilians, and the number of cases where not only experts but also non-experts use UAVs is increasing. Autonomous UAVs are used in various fields such as logistics, delivery, and delivery, but in areas where GPS is not available or external forces cannot be estimated, resulting in accidents such as crashes and falls. In this research, the method is proposed as a model for predicting the path and tilt of UAVs using ANN and predicting acceleration and thus predicting paths for new paths to control autonomous flighting. As a result, the position of the UAV was predicted with high accuracy, and the direction of inclination was predicted with high accuracy. In addition, path prediction for the new path was also predicted with high accuracy. This result shows that it can provide high accuracy of ANN and be the basis for positive decision support for pilots and airborne autonomous navigation systems.