AI RESEARCH
Towards Generative Predictive Display for Vision-Based Teleoperation: A Zero-Shot Benchmark of Off-the-Shelf Video Models
arXiv CS.CV
•
ArXi:2605.09670v1 Announce Type: cross Teleoperation systems are fundamentally limited by communication latency, which degrades situational awareness and control performance. Predictive display aims to mitigate this limitation by presenting an estimate of the current visual state rather than delayed observations. While recent advances in generative video models enable high-quality video synthesis, their suitability for latency-sensitive predictive display remains unclear.