Sep 2026
Learning Reliable Parking Policies via Offline Reinforcement Learning with Quantized Action Representations
This work develops a waypoint-level offline reinforcement learning framework for interaction-aware autonomous parking, which attains the highest parking success rate among all baselines and transfers reliably to unseen parking slots.
Ze-Wei Yang, Zeng-Qi Peng, Jun Ma
· arXiv.org · 0 citations