Visual Homing in Outdoor Robots Using Mushroom Body Circuits and Learning Walks

Ants achieve robust visual homing with minimal sensory input and only a few learning walks, inspiring biomimetic solu tions for autonomous navigation. While Mushroom Body (MB) models hav e been used in robotic route following, they have not yet been appli ed to visual homing. We present the first real-world implementation of a l ateralized MB architecture for visual homing onboard a compact autonom ous car-like robot. We test whether the sign of the angular path integ ration (PI) signal can categorize panoramic views, acquired during lea rning walks and encoded in the MB, into "goal on the left" and "goal on the r ight" memory banks, enabling robust homing in natural outdoor set tings. We validate this approach through four incremental experimen ts: (1) simulation showing attractor-like nest dynamics; (2) real-wor ld homing after decoupled learning walks, producing nest search behavior; (3) homing after random walks using noisy PI emulated with GPS-RTK; and (4) precise stopping-at-the-goal behavior enabled by a fifth MB Output Neuron (MBON) encoding goal-views to control velocity. This mi mics the accurate homing behavior of ants and functionally resemble s waypoint-based position control in robotics, despite relying solely on visual input. Operating at 8 Hz on a Raspberry Pi 4 with 32 32 pixel views and a memory footprint under 9 kB, our system offers a biologically grounded, resource-efficient solution for autonomous visual homing.