FocalAD: Local Motion Planning for End-to-End Autonomous Driving

2025-06-13

AI Summary

In end-to-end autonomous driving, motion prediction often relies on global features while neglecting locally critical interacting agents—those most influential to ego-vehicle planning—leading to insufficient risk awareness and degraded planning reliability. To address this, we propose a novel motion modeling paradigm centered on focal local interactions. Specifically, we design the Ego-Local-Agents Interactor (ELAI), a graph neural module incorporating localized attention to dynamically weight neighboring agents’ motions. We further introduce the Focal-Local-Agents Loss (FLA Loss) to reinforce gradient backpropagation for these key interactions. Integrated into a fully differentiable end-to-end planning architecture, our method achieves state-of-the-art performance on nuScenes and Bench2Drive. Under adversarial evaluation on Adv-nuScenes, it reduces collision rates by 41.9% versus DiffusionDrive and by 15.6% versus SparseDrive, demonstrating substantial improvements in planning safety and robustness.

Technology Category

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

Application Category

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

Abstract

In end-to-end autonomous driving,the motion prediction plays a pivotal role in ego-vehicle planning. However, existing methods often rely on globally aggregated motion features, ignoring the fact that planning decisions are primarily influenced by a small number of locally interacting agents. Failing to attend to these critical local interactions can obscure potential risks and undermine planning reliability. In this work, we propose FocalAD, a novel end-to-end autonomous driving framework that focuses on critical local neighbors and refines planning by enhancing local motion representations. Specifically, FocalAD comprises two core modules: the Ego-Local-Agents Interactor (ELAI) and the Focal-Local-Agents Loss (FLA Loss). ELAI conducts a graph-based ego-centric interaction representation that captures motion dynamics with local neighbors to enhance both ego planning and agent motion queries. FLA Loss increases the weights of decision-critical neighboring agents, guiding the model to prioritize those more relevant to planning. Extensive experiments show that FocalAD outperforms existing state-of-the-art methods on the open-loop nuScenes datasets and closed-loop Bench2Drive benchmark. Notably, on the robustness-focused Adv-nuScenes dataset, FocalAD achieves even greater improvements, reducing the average colilision rate by 41.9% compared to DiffusionDrive and by 15.6% compared to SparseDrive.

Problem

• Improves ego-vehicle planning by focusing on local interactions
• Enhances motion prediction with critical local neighbor attention
• Reduces collision rates in autonomous driving scenarios

Innovation

• Focuses on critical local neighbor interactions
• Uses graph-based ego-centric interaction representation
• Increases weights of decision-critical neighboring agents