A two-stage pipeline with an attention-augmented gated recurrent wind estimator and proximal policy optimization controller reduces quadrotor horizontal trajectory tracking error by 48% versus a wind-blind baseline in simulated von Karman turbulence.
Geometric Adaptive Control With Neural Networks for a Quadrotor in Wind Fields,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Wind-Aware Reinforcement Learning Control of a Small Quadrotor Using Learned Onboard Wind Estimation in Simulated Atmospheric Turbulence
A two-stage pipeline with an attention-augmented gated recurrent wind estimator and proximal policy optimization controller reduces quadrotor horizontal trajectory tracking error by 48% versus a wind-blind baseline in simulated von Karman turbulence.