Beyond Patterns: Harnessing Causal Logic for Autonomous Driving Trajectory Prediction

2025-05-11

AI Summary

Current trajectory prediction methods for autonomous driving heavily rely on statistical correlations, resulting in poor robustness and limited generalizability. To address this, this paper introduces causal inference—its first application to this task—proposing a spatiotemporal disentanglement and progressive multimodal fusion framework grounded in causal graph modeling. The method explicitly encodes causal mechanisms among traffic agents, eliminates spurious correlations, and integrates counterfactual reasoning with a lightweight real-time inference engine. Evaluated on five real-world benchmarks—including ApolloScape and nuScenes—it achieves state-of-the-art performance: RMSE improves by 12.7% and final displacement error (FDE) by 14.3%, while satisfying real-time constraints. The core contribution is the establishment of the first interpretable and generalizable causal trajectory prediction paradigm, overcoming fundamental limitations of conventional correlation-based modeling.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate trajectory prediction has long been a major challenge for autonomous driving (AD). Traditional data-driven models predominantly rely on statistical correlations, often overlooking the causal relationships that govern traffic behavior. In this paper, we introduce a novel trajectory prediction framework that leverages causal inference to enhance predictive robustness, generalization, and accuracy. By decomposing the environment into spatial and temporal components, our approach identifies and mitigates spurious correlations, uncovering genuine causal relationships. We also employ a progressive fusion strategy to integrate multimodal information, simulating human-like reasoning processes and enabling real-time inference. Evaluations on five real-world datasets--ApolloScape, nuScenes, NGSIM, HighD, and MoCAD--demonstrate our model's superiority over existing state-of-the-art (SOTA) methods, with improvements in key metrics such as RMSE and FDE. Our findings highlight the potential of causal reasoning to transform trajectory prediction, paving the way for robust AD systems.

Problem

• Enhancing autonomous driving trajectory prediction accuracy
• Addressing spurious correlations with causal inference
• Integrating multimodal data for human-like reasoning

Innovation

• Leverages causal inference for robust trajectory prediction
• Decomposes environment into spatial and temporal components
• Employs progressive fusion for multimodal information integration