Learning to Fly in Seconds

arXiv:2311.13081v2 Announce Type: replace-cross
Abstract: Learning-based methods, particularly Reinforcement Learning (RL), hold great promise for streamlining deployment, enhancing performance, and achieving generalization in the control of autonomous multirotor aerial vehicles. Deep RL has been able to control complex systems with impressive fidelity and agility in simulation but the simulation-to-reality transfer often brings a hard-to-bridge reality gap. Moreover, RL is commonly plagued by prohibitively long training times. In this work, we propose a novel asymmetric actor-critic-based architecture coupled with a highly …

abstract aerial agility arxiv autonomous bridge complex systems control cs.ai cs.lg cs.ro cs.sy deep rl deployment eess.sy fidelity fly performance reality reinforcement reinforcement learning simulation systems the simulation transfer type vehicles