Preprint
Aug 2026
RedLight-VLA: Models for traffic-rule grounding and behavioral emphasis in driving policies
RedLight-VLA, a training objective that uses expert futures and automatically generated perception targets without additional manual rule annotation, is presented, which emphasizes rare deceleration and acceleration using rotation-invariant longitudinal dynamics and a scale-preserving reduction that exactly recovers th...
B. Sudhakar, S. Sridhar, Sandipan Das et al.
· 0 citations