analiza aiqbench
This paper applies a proprioceptive recurrent policy, trained with PPO in simulation, to recover quadrotors after collisions without position, velocity, or heading estimates. The hardware results (19 of 22 recoveries versus 1 of 14 for the stock autopilot) are notable, but they come from one platform and a specific wall-impact setup, so generalization is unproven. For AI readers, the useful takeaway is the sim-to-real pipeline and the trust-gate design for handing control back to a conventional estimator.
REFLEX: A Learned On-Board Reflex for Robust Autonomous Collision Recovery of Soft Quadrotors
arXiv:2610.15867v1 Announce Type: new Abstract: Collision-resilient airframes allow quadrotors to survive impacts, but do not ensure autonomous post-collision recovery. Innovation gating in stock onboard estimators often rejects valid external velocity measurements after high-speed impacts, compromising recovery. We present REFLEX, a collision-recovery flight mode that runs entirely on the flight controller, requiring neither contact sensors nor a companion computer. Its proprioceptive recurrent policy stabilizes the vehicle without position, velocity, heading, or altitude estimates. The translational estimate is reinitialized while REFLEX is active, and a trust gate checks estimator consistency before returning control to the stock autopilot. The policy is trained in simulation using proximal policy optimization, with episodes initialized from post-impact states measured in 47 real collisions. In hardware experiments, REFLEX recovered 19 of 22 wall impacts at 5.03 $\pm$ 0.09 m/s, compared with 1 of 14 for the matched stock autopilot. These included 15 consecutive recoveries in a single flight. A separate exploratory flight demonstrated two additional recoveries at 6.02 and 6.05 m/s within the same flight, exceeding the previously reported velocity for autonomous wall-collision recovery by 64%.
Abstract:Collision-resilient airframes allow quadrotors to survive impacts, but do not ensure autonomous post-collision recovery. Innovation gating in stock onboard estimators often rejects valid external velocity measurements after high-speed impacts, compromising recovery. We present REFLEX, a collision-recovery flight mode that runs entirely on the flight controller, requiring neither contact sensors nor a companion computer. Its proprioceptive recurrent policy stabilizes the vehicle without position, velocity, heading, or altitude estimates. The translational estimate is reinitialized while REFLEX is active, and a trust gate checks estimator consistency before returning control to the stock autopilot. The policy is trained in simulation using proximal policy optimization, with episodes initialized from post-impact states measured in 47 real collisions. In hardware experiments, REFLEX recovered 19 of 22 wall impacts at 5.03 $\pm$ 0.09 m/s, compared with 1 of 14 for the matched stock autopilot. These included 15 consecutive recoveries in a single flight. A separate exploratory flight demonstrated two additional recoveries at 6.02 and 6.05 m/s within the same flight, exceeding the previously reported velocity for autonomous wall-collision recovery by 64%.
J.H. and L.G. contributed equally;S.L. and S.M. jointly advised;Supplementary video: this https URL by EU Horizon Europe project SPEAR (grant no. 101119774);Submitted to ICRA 2027
Robotics (cs.RO)
arXiv:2610.15867 [cs.RO]
(or arXiv:2610.15867v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.15867 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 215 cuvinte
Subiecte conexe
LAEI: Layered Autonomous Edge Intelligence Framework for Robust UAV Swarm Operations
arXiv · 9 iun. 2026
The Machines that Make the Machines
NVIDIA · 7 oct. 2026
Agile Fall Recovery for Quadrotors with Bidirectional Thrust via Reinforcement Learning
arXiv · 16 iun. 2026
Synthetic-to-Real Pipeline for Safe Landing Zone Detection
arXiv · 16 iun. 2026