GoalFlow: Goal-Driven Flow Matching for Multimodal Trajectories Generation in End-to-End Autonomous Driving

2025-03-07

AI Summary

End-to-end autonomous driving suffers from severe modality divergence, degraded trajectory quality, and inconsistency between guidance signals (e.g., goal points) and scene context in multimodal trajectory generation. Method: We propose a goal-point-driven flow matching framework: (i) the goal point is explicitly incorporated as a conditional input into the flow matching process for the first time; (ii) a scene-aware goal-point scoring network and trajectory reweighting & selection mechanism are designed to ensure high-quality, semantically consistent multimodal outputs; (iii) single-step denoising generation is supported, drastically reducing computational overhead. Contribution/Results: On the NuScenes-based Navsim benchmark, our method achieves a PDMS score of 90.3, establishing new state-of-the-art performance. Notably, its single-step inference already surpasses the multi-step results of existing diffusion-based approaches, with significantly accelerated runtime.

Technology Category

• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We propose GoalFlow, an end-to-end autonomous driving method for generating high-quality multimodal trajectories. In autonomous driving scenarios, there is rarely a single suitable trajectory. Recent methods have increasingly focused on modeling multimodal trajectory distributions. However, they suffer from trajectory selection complexity and reduced trajectory quality due to high trajectory divergence and inconsistencies between guidance and scene information. To address these issues, we introduce GoalFlow, a novel method that effectively constrains the generative process to produce high-quality, multimodal trajectories. To resolve the trajectory divergence problem inherent in diffusion-based methods, GoalFlow constrains the generated trajectories by introducing a goal point. GoalFlow establishes a novel scoring mechanism that selects the most appropriate goal point from the candidate points based on scene information. Furthermore, GoalFlow employs an efficient generative method, Flow Matching, to generate multimodal trajectories, and incorporates a refined scoring mechanism to select the optimal trajectory from the candidates. Our experimental results, validated on the Navsimcite{Dauner2024_navsim}, demonstrate that GoalFlow achieves state-of-the-art performance, delivering robust multimodal trajectories for autonomous driving. GoalFlow achieved PDMS of 90.3, significantly surpassing other methods. Compared with other diffusion-policy-based methods, our approach requires only a single denoising step to obtain excellent performance. The code is available at https://github.com/YvanYin/GoalFlow.

Problem

• Generates high-quality multimodal trajectories for autonomous driving.
• Reduces trajectory divergence and improves consistency with scene information.
• Selects optimal trajectories efficiently using a novel scoring mechanism.

Innovation

• GoalFlow introduces goal point constraints
• Novel scoring mechanism selects optimal goal
• Flow Matching generates multimodal trajectories efficiently