RealDrive: Retrieval-Augmented Driving with Diffusion Models

TL;DR

RealDrive introduces a retrieval-augmented diffusion planning framework for autonomous driving, enhancing safety, control, and generalization by leveraging expert demonstrations and task-relevant retrieval.

Contribution

The paper presents a novel retrieval-augmented diffusion planning method that improves safety and diversity in autonomous driving behaviors.

Findings

40% reduction in collision rate on Waymo dataset

Improved generalization to rare scenarios

Enhanced trajectory diversity

Abstract

Learning-based planners generate natural human-like driving behaviors by learning to reason about nuanced interactions from data, overcoming the rigid behaviors that arise from rule-based planners. Nonetheless, data-driven approaches often struggle with rare, safety-critical scenarios and offer limited controllability over the generated trajectories. To address these challenges, we propose RealDrive, a Retrieval-Augmented Generation (RAG) framework that initializes a diffusion-based planning policy by retrieving the most relevant expert demonstrations from the training dataset. By interpolating between current observations and retrieved examples through a denoising process, our approach enables fine-grained control and safe behavior across diverse scenarios, leveraging the strong prior provided by the retrieved scenario. Another key insight we produce is that a task-relevant retrieval model trained with planning-based objectives results in superior planning performance in our framework compared to a task-agnostic retriever. Experimental results demonstrate improved generalization to long-tail events and enhanced trajectory diversity compared to standard learning-based planners -- we observe a 40% reduction in collision rate on the Waymo Open Motion dataset with RAG.