In First-Person View (FPV) drone flying, professional pilots demonstrate remarkable skill, navigating through complex environments with precision and flair. The essence of FPV flight lies not just in efficiency or speed, but in the "cool" factor — the ability to perform dynamic, agile maneuvers that captivate and impress.
This project explores the challenge of capturing this "coolness" factor in optimization, enabling the development of an autonomous flight system capable of replicating the nuanced flight patterns of expert human pilots. Our research focuses on formulating these advanced maneuvers and implementing them through a vision-based system, allowing drones to autonomously navigate through cluttered spaces like forests with the same level of skill and style as their human counterparts.
In First-Person View (FPV) drone flying, professional pilots demonstrate remarkable skill, navigating through complex environments with precision and flair. The essence of FPV flight lies not just in efficiency or speed, but in the "cool" factor — the ability to perform dynamic, agile maneuvers that captivate and impress.
This project explores the challenge of capturing this "coolness" factor in optimization, enabling the development of an autonomous flight system capable of replicating the nuanced flight patterns of expert human pilots. Our research focuses on formulating these advanced maneuvers and implementing them through a vision-based system, allowing drones to autonomously navigate through cluttered spaces like forests with the same level of skill and style as their human counterparts.
To create a sophisticated autonomous FPV flight system that integrates advanced computer vision and control algorithms, enabling drones to autonomously execute complex, human-like maneuvers in cluttered and dynamically changing environments.
To create a sophisticated autonomous FPV flight system that integrates advanced computer vision and control algorithms, enabling drones to autonomously execute complex, human-like maneuvers in cluttered and dynamically changing environments.
Yunlong Song (song@ifi.uzh.ch), Nico Messikommer (nmessi@ifi.uzh.ch)
Yunlong Song (song@ifi.uzh.ch), Nico Messikommer (nmessi@ifi.uzh.ch)