RAD-2: Scaling Reinforcement Learning in a Generator-Discriminator Framework

TL;DR

RAD-2 introduces a generator-discriminator framework combining diffusion models and reinforcement learning to improve autonomous driving planning, achieving higher safety and smoother driving in urban environments.

Contribution

The paper presents a novel unified generator-discriminator approach with new RL techniques and a high-throughput simulation environment for scalable, stable closed-loop autonomous driving planning.

Findings

Reduced collision rate by 56% compared to diffusion-based planners

Improved perceived safety and driving smoothness in real-world urban traffic

Enhanced training stability through decoupled trajectory generation and evaluation

Abstract

High-level autonomous driving requires motion planners capable of modeling multimodal future uncertainties while remaining robust in closed-loop interactions. Although diffusion-based planners are effective at modeling complex trajectory distributions, they often suffer from stochastic instabilities and the lack of corrective negative feedback when trained purely with imitation learning. To address these issues, we propose RAD-2, a unified generator-discriminator framework for closed-loop planning. Specifically, a diffusion-based generator is used to produce diverse trajectory candidates, while an RL-optimized discriminator reranks these candidates according to their long-term driving quality. This decoupled design avoids directly applying sparse scalar rewards to the full high-dimensional trajectory space, thereby improving optimization stability. To further enhance reinforcement learning, we introduce Temporally Consistent Group Relative Policy Optimization, which exploits temporal coherence to alleviate the credit assignment problem. In addition, we propose On-policy Generator Optimization, which converts closed-loop feedback into structured longitudinal optimization signals and progressively shifts the generator toward high-reward trajectory manifolds. To support efficient large-scale training, we introduce BEV-Warp, a high-throughput simulation environment that performs closed-loop evaluation directly in Bird's-Eye View feature space via spatial warping. RAD-2 reduces the collision rate by 56% compared with strong diffusion-based planners. Real-world deployment further demonstrates improved perceived safety and driving smoothness in complex urban traffic.