DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning

2025-06-07

AI Summary

To address insufficient precision in multi-candidate trajectory selection for end-to-end autonomous driving planning—particularly weak discriminative capability in long-tail and out-of-distribution (OOD) scenarios—this paper proposes a safety-oriented trajectory selection framework grounded in a selective paradigm. Our method introduces three key innovations: (1) a coarse-to-fine progressive candidate filtering mechanism to improve optimal trajectory retrieval efficiency; (2) rotation-based data augmentation to enhance OOD robustness; and (3) a self-distillation-driven contrastive learning framework that jointly leverages multi-scale trajectory scoring and geometry-aware data augmentation to stabilize training and strengthen safety-critical semantic representations. Evaluated on the NAVSIM v1/v2 benchmarks, our approach achieves 93.5% PDMS and 87.1% EPDMS, significantly improving collision avoidance and traffic-rule compliance—without requiring additional annotated data.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Natural Language Processing: Safety and Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In complex driving environments, autonomous vehicles must navigate safely. Relying on a single predicted path, as in regression-based approaches, usually does not explicitly assess the safety of the predicted trajectory. Selection-based methods address this by generating and scoring multiple trajectory candidates and predicting the safety score for each, but face optimization challenges in precisely selecting the best option from thousands of possibilities and distinguishing subtle but safety-critical differences, especially in rare or underrepresented scenarios. We propose DriveSuprim to overcome these challenges and advance the selection-based paradigm through a coarse-to-fine paradigm for progressive candidate filtering, a rotation-based augmentation method to improve robustness in out-of-distribution scenarios, and a self-distillation framework to stabilize training. DriveSuprim achieves state-of-the-art performance, reaching 93.5% PDMS in NAVSIM v1 and 87.1% EPDMS in NAVSIM v2 without extra data, demonstrating superior safetycritical capabilities, including collision avoidance and compliance with rules, while maintaining high trajectory quality in various driving scenarios.

Problem

• Improving trajectory selection safety in autonomous driving
• Optimizing selection from thousands of trajectory candidates
• Enhancing robustness in rare driving scenarios

Innovation

• Coarse-to-fine progressive candidate filtering
• Rotation-based augmentation for robustness
• Self-distillation framework stabilizes training