CogDrive: Cognition-Driven Multimodal Prediction-Planning Fusion for Safe Autonomy

2025-12-02

AI Summary

To address the challenges of modeling rare interactive behaviors and ensuring adaptive safety planning in mixed traffic, this paper proposes a cognition-driven multimodal prediction and planning integration framework. Methodologically, it introduces: (i) topology-aware motion semantics for cognitive interaction representation, explicitly encoding sparse critical behaviors; (ii) differentiable modality-specific losses combined with an emergency response mechanism to enhance interpretability in sparse-behavior learning and re-planning safety; and (iii) tight coupling of prediction and decision-making via nearest-neighbor relational encoding, multimodal Gaussian decoding, safety-stable trajectory optimization, and short-horizon consistent branching planning. Evaluated on Argoverse2 and INTERACTION datasets, the framework achieves significant improvements in trajectory accuracy and low false-negative rates. Closed-loop simulations further demonstrate its adaptive safe navigation capability in complex scenarios such as highway on-ramps and intersections.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Computer Vision: Multi-modal Vision

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Safe autonomous driving in mixed traffic requires a unified understanding of multimodal interactions and dynamic planning under uncertainty. Existing learning based approaches struggle to capture rare but safety critical behaviors, while rule based systems often lack adaptability in complex interactions. To address these limitations, CogDrive introduces a cognition driven multimodal prediction and planning framework that integrates explicit modal reasoning with safety aware trajectory optimization. The prediction module adopts cognitive representations of interaction modes based on topological motion semantics and nearest neighbor relational encoding. With a differentiable modal loss and multimodal Gaussian decoding, CogDrive learns sparse and unbalanced interaction behaviors and improves long horizon trajectory prediction. The planning module incorporates an emergency response concept and optimizes safety stabilized trajectories, where short term consistent branches ensure safety during replanning cycles and long term branches support smooth and collision free motion under low probability switching modes. Experiments on Argoverse2 and INTERACTION datasets show that CogDrive achieves strong performance in trajectory accuracy and miss rate, while closed loop simulations confirm adaptive behavior in merge and intersection scenarios. By combining cognitive multimodal prediction with safety oriented planning, CogDrive offers an interpretable and reliable paradigm for safe autonomy in complex traffic.

Problem

• Enhances multimodal prediction for rare safety-critical behaviors
• Integrates cognitive reasoning with safety-aware trajectory optimization
• Improves adaptability in complex traffic interactions and planning

Innovation

• Cognition-driven multimodal prediction and planning fusion
• Differentiable modal loss with multimodal Gaussian decoding
• Safety-stabilized trajectory optimization with emergency response