RECON: Replayability-Enhanced Cloning for Off-Trajectory Navigation

A digital copy can look convincing and still give a robot the wrong map. This matters when a drone learns in a 3D clone built from a short scouting flight, then flies beyond the views used to build it. We study task replayability, the geometric reliability of a clone along later control routes. RECON represents this reliability with a Replayability Field $E:\mathcal{P}\to [0,1]$ computed from refined Gaussians and scout poses with no learned parameters. On 3000 test-world poses, $E$ has $-0.58$ Spearman correlation with depth error and detects high-gap geometry with $0.83$ AUROC. The policy reads Gaussian tokens and $E$, while a support cost shapes its training routes. With architecture, source worlds, training budget, optimizer and seeds fixed, these components raise closed-loop success from $0.44$ to $0.52$ and reduce collisions from $0.56$ to $0.48$ on 600 episodes across 6 unseen reconstructed worlds. Across 40 real-UAV flights in 4 environments, RECON reaches $0.70$ success, compared with $0.65$ for the strongest image baseline.