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

TL;DR

This paper introduces a physics-guided causal model that improves zero-shot trajectory prediction in autonomous driving by extracting domain-invariant features and integrating them with kinematic models, demonstrating superior generalization across unseen cities.

Contribution

The paper proposes a novel PCM framework with a disentangled scene encoder and causal ODE decoder for better zero-shot generalization in trajectory prediction.

Findings

Outperforms baselines in unseen cities

Effective extraction of domain-invariant features

Enhanced integration of scene context and kinematics

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.