REAP: Reinforcement-Learning End-to-End Autonomous Parking with Gaussian Splatting Simulator for Real2Sim2Real Transfer

2026-05-09

AI Summary

This work addresses the challenge of coordinating perception and planning in traditional multi-stage automated parking systems, which often fail in extreme scenarios such as mechanical parking spots or dead-end alleys. The authors propose REAP, an end-to-end reinforcement learning framework that, for the first time, leverages 3D Gaussian Splatting to construct a Real2Sim2Real simulation system, enabling efficient transfer from real-world data to simulation and back. By integrating asymmetric reinforcement learning, behavior cloning for accelerated training, a soft predictive collision penalty mechanism, and knowledge distillation from a rule-based planner, REAP significantly improves parking success rates in highly constrained environments. Experimental results demonstrate that the method trains efficiently in simulation and successfully executes diverse complex parking maneuvers on a real vehicle, validating the feasibility of end-to-end reinforcement learning for ultra-narrow mechanical parking tasks.

Abstract

In recent years, autonomous parking has made significant advances, yet parking tasks still face challenges in extreme scenarios such as mechanical and dead-end parking slots, often resulting in failures. This is mainly due to traditional parking methods adopting a multistage approach, lacking the ability to optimize the parking problem as a whole. End-to-end methods enable joint optimization across perception and planning modules to eliminate the accumulation of errors, enhancing algorithm performance in extreme scenarios. Although several end-to-end parking methods use imitation or reinforcement learning, the former is limited by data cost and distribution coverage, while the latter suffers from inefficient exploration. To address these challenges, we propose a Reinforcement learning End-to-end Autonomous Parking method (REAP). REAP employs Soft Actor-Critic (SAC) within an asymmetric reinforcement learning framework to improve training efficiency and inference performance. To accelerate model convergence, we distill the capabilities of a rule-based planner into the end-to-end network through behavior cloning. We further introduce a soft predictive collision penalty mechanism to reduce collision rates by penalizing obstacle-approaching actions. To ensure that the trained reinforcement learning network can directly transfer to real-world scenarios, we have established a Real2Sim2Real simulator. In the Real2Sim step, we use 3D Gaussian Splatting (3DGS) to transform real-world scenes into digital scenes. In the Sim2Real step, we deploy the end-to-end model onto the vehicle to bridge the Sim2Real gap. Trained in the 3DGS simulator and deployed on physical vehicles, REAP successfully parks in various types of parking spaces, especially demonstrating the feasibility of end-to-end RL parking in extremely narrow mechanical slots.

Problem

• autonomous parking
• extreme scenarios
• end-to-end learning
• real2sim2real transfer
• reinforcement learning

Innovation

• end-to-end reinforcement learning
• 3D Gaussian Splatting
• Real2Sim2Real transfer
• autonomous parking
• behavior cloning