A Generalizable Physics-guided Causal Model for Trajectory Prediction in Autonomous Driving

2026-02-15

AI Summary

This work addresses the limited zero-shot generalization capability of autonomous driving trajectory prediction models in unseen scenarios by proposing a physics-guided causal framework that integrates domain-invariant kinematic priors with data-driven learning to enhance cross-domain generalization. The core of the approach comprises an intervention-disentangled scene encoder that extracts domain-invariant features and a decoder that combines a causal attention mechanism with ordinary differential equations (CausalODE) to generate physically plausible trajectories. Experimental results demonstrate that the proposed method significantly outperforms existing baselines across multiple real-world autonomous driving datasets and exhibits exceptional zero-shot generalization performance in previously unseen urban environments.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Machine Learning: Causal Learning
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• 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

Trajectory prediction for traffic agents is critical for safe autonomous driving. However, achieving effective zero-shot generalization in previously unseen domains remains a significant challenge. Motivated by the consistent nature of kinematics across diverse domains, we aim to incorporate domain-invariant knowledge to enhance zero-shot trajectory prediction capabilities. The key challenges include: 1) effectively extracting domain-invariant scene representations, and 2) integrating invariant features with kinematic models to enable generalized predictions. To address these challenges, we propose a novel generalizable Physics-guided Causal Model (PCM), which comprises two core components: a Disentangled Scene Encoder, which adopts intervention-based disentanglement to extract domain-invariant features from scenes, and a CausalODE Decoder, which employs a causal attention mechanism to effectively integrate kinematic models with meaningful contextual information. Extensive experiments on real-world autonomous driving datasets demonstrate our method's superior zero-shot generalization performance in unseen cities, significantly outperforming competitive baselines. The source code is released at https://github.com/ZY-Zong/Physics-guided-Causal-Model.

Problem

• trajectory prediction
• zero-shot generalization
• autonomous driving
• domain-invariant representation
• kinematic modeling

Innovation

• Physics-guided
• Causal Model
• Zero-shot Generalization
• Disentangled Representation
• Trajectory Prediction