ReVNM: Learning-Based Visual Navigation from a Remote Camera

Michikuni Eguchi, Kohei Honda, Masafumi Endo +2

Visual Navigation Models (VNMs) enable robots to navigate from egocentric visual observations without geometric localization and planning, but long-range navigation still requires…