The author introduces Lander.AI, an advanced Deep Reinforcement Learning agent designed to enhance drone autonomy and safety during dynamic platform landings. The approach leverages bio-inspired learning to adapt to external forces like wind, showcasing significant improvements in landing precision and error recovery.
Lander.AIは風の影響を受ける状況で高精度な着陸を実現し、ドローンの自律性と安全性を向上させる。
This research paper proposes a novel vision-based system for assessing the risk of autonomous drone landings in complex urban environments, leveraging semantic segmentation with deep learning to identify safe landing zones and enhance drone safety in populated areas.