Skip to content

Translation in progress Some entries still include text in the original language.

Failure-First·Indexed by this site · Originally published

无人机敏捷飞行强化学习:兼顾竞速性能与泛化

[Daily Paper] Bridging Performance and Generalization in Reinforcement Learning for Agile Flight

AI summary

Green 等人提出一种强化学习框架,让无人机在未见过赛道上零样本泛化飞行,同时保持竞速速度。框架跨多条赛道并行采集 rollout 训练,用 B-Spline 生成几何可行赛道,并自动替换奖励饱和的任务。未见赛道上成功率达 69.8%,性能加权得分 0.1652,泛化较此前最优方法提升 7.4 倍,速度比 EaP 快 37.73%。

Read originalfailurefirst.org