Toward Physically Consistent Driving Video World Models under Challenging Trajectories

TL;DR

PhyGenesis is a novel world model for autonomous driving videos that ensures physical consistency and high visual fidelity, especially under challenging or counterfactual trajectories, by using a physics-rich dataset and specialized components.

Contribution

The paper introduces PhyGenesis, a framework with a physical condition generator and physics-enhanced video generator trained on a diverse dataset, to produce physically consistent driving videos under challenging scenarios.

Findings

Outperforms state-of-the-art methods on challenging trajectories.

Successfully generates high-fidelity, physically consistent driving videos.

Enables trajectory correction for physically implausible inputs.

Abstract

Video generation models have shown strong potential as world models for autonomous driving simulation. However, existing approaches are primarily trained on real-world driving datasets, which mostly contain natural and safe driving scenarios. As a result, current models often fail when conditioned on challenging or counterfactual trajectories-such as imperfect trajectories generated by simulators or planning systems-producing videos with severe physical inconsistencies and artifacts. To address this limitation, we propose PhyGenesis, a world model designed to generate driving videos with high visual fidelity and strong physical consistency. Our framework consists of two key components: (1) a physical condition generator that transforms potentially invalid trajectory inputs into physically plausible conditions, and (2) a physics-enhanced video generator that produces high-fidelity multi-view driving videos under these conditions. To effectively train these components, we construct a large-scale, physics-rich heterogeneous dataset. Specifically, in addition to real-world driving videos, we generate diverse challenging driving scenarios using the CARLA simulator, from which we derive supervision signals that guide the model to learn physically grounded dynamics under extreme conditions. This challenging-trajectory learning strategy enables trajectory correction and promotes physically consistent video generation. Extensive experiments demonstrate that PhyGenesis consistently outperforms state-of-the-art methods, especially on challenging trajectories. Our project page is available at: https://wm-research.github.io/PhyGenesis/.