Flow Matching-Based Autonomous Driving Planning with Advanced Interactive Behavior Modeling

2025-10-13

AI Summary

Imitation learning for autonomous driving is hindered by the difficulty of modeling complex interactive behaviors and the scarcity of high-quality interaction data. Method: We propose Flow Planner, a novel planning framework comprising (i) a trajectory segmentation tokenization scheme for fine-grained motion representation; (ii) a spatiotemporal fusion Transformer architecture that explicitly models multi-agent dynamic interactions; and (iii) a classifier-free guided flow matching generative framework enabling efficient, stable, and multimodal interaction-aware planning. Flow Planner dynamically modulates inter-agent interaction weights to enhance planning consistency and scene adaptability. Results: Flow Planner achieves state-of-the-art performance among learning-based methods on nuPlan and the high-interaction-density interPlan benchmark, demonstrating significant improvements in high-value interactive scenarios—particularly lane-change games and unmarked intersection navigation.

Technology Category

• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Planning with Language Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search

Abstract

Modeling interactive driving behaviors in complex scenarios remains a fundamental challenge for autonomous driving planning. Learning-based approaches attempt to address this challenge with advanced generative models, removing the dependency on over-engineered architectures for representation fusion. However, brute-force implementation by simply stacking transformer blocks lacks a dedicated mechanism for modeling interactive behaviors that are common in real driving scenarios. The scarcity of interactive driving data further exacerbates this problem, leaving conventional imitation learning methods ill-equipped to capture high-value interactive behaviors. We propose Flow Planner, which tackles these problems through coordinated innovations in data modeling, model architecture, and learning scheme. Specifically, we first introduce fine-grained trajectory tokenization, which decomposes the trajectory into overlapping segments to decrease the complexity of whole trajectory modeling. With a sophisticatedly designed architecture, we achieve efficient temporal and spatial fusion of planning and scene information, to better capture interactive behaviors. In addition, the framework incorporates flow matching with classifier-free guidance for multi-modal behavior generation, which dynamically reweights agent interactions during inference to maintain coherent response strategies, providing a critical boost for interactive scenario understanding. Experimental results on the large-scale nuPlan dataset and challenging interactive interPlan dataset demonstrate that Flow Planner achieves state-of-the-art performance among learning-based approaches while effectively modeling interactive behaviors in complex driving scenarios.

Problem

• Modeling interactive driving behaviors in complex scenarios
• Addressing scarcity of interactive driving data for learning
• Improving multi-modal behavior generation with dynamic interactions

Innovation

• Flow matching with classifier-free guidance for generation
• Fine-grained trajectory tokenization to reduce complexity
• Efficient spatiotemporal fusion architecture for interaction modeling