Bootstrapping Reinforcement Learning with Imitation for Vision-Based Agile Flight

arXiv:2403.12203v1 Announce Type: cross
Abstract: We combine the effectiveness of Reinforcement Learning (RL) and the efficiency of Imitation Learning (IL) in the context of vision-based, autonomous drone racing. We focus on directly processing visual input without explicit state estimation. While RL offers a general framework for learning complex controllers through trial and error, it faces challenges regarding sample efficiency and computational demands due to the high dimensionality of visual inputs. Conversely, IL demonstrates efficiency in learning from visual demonstrations but …

abstract agile arxiv autonomous bootstrapping context cs.cv cs.lg cs.ro drone drone racing efficiency focus framework general imitation learning processing racing reinforcement reinforcement learning state through type vision visual