Vehicle Dynamics Embedded World Models for Autonomous Driving

TL;DR

This paper introduces VDD, a novel world model for autonomous driving that separates vehicle and environment dynamics, enhancing robustness and generalization across diverse vehicle types.

Contribution

The paper proposes the Vehicle Dynamics embedded Dreamer (VDD), which decouples ego-vehicle and environmental dynamics modeling, and introduces PAD and PAT strategies to improve robustness.

Findings

VDD outperforms existing models in simulated driving tasks.

Decoupling dynamics improves generalization across vehicle types.

PAD and PAT strategies enhance policy robustness.

Abstract

World models have gained significant attention as a promising approach for autonomous driving. By emulating human-like perception and decision-making processes, these models can predict and adapt to dynamic environments. Existing methods typically map high-dimensional observations into compact latent spaces and learn optimal policies within these latent representations. However, prior work usually jointly learns ego-vehicle dynamics and environmental transition dynamics from the image input, leading to inefficiencies and a lack of robustness to variations in vehicle dynamics. To address these issues, we propose the Vehicle Dynamics embedded Dreamer (VDD) method, which decouples the modeling of ego-vehicle dynamics from environmental transition dynamics. This separation allows the world model to generalize effectively across vehicles with diverse parameters. Additionally, we introduce two strategies to further enhance the robustness of the learned policy: Policy Adjustment during Deployment (PAD) and Policy Augmentation during Training (PAT). Comprehensive experiments in simulated environments demonstrate that the proposed model significantly improves both driving performance and robustness to variations in vehicle dynamics, outperforming existing approaches.