InDRiVE: Reward-Free World-Model Pretraining for Autonomous Driving via Latent Disagreement

TL;DR

This paper introduces InDRiVE, a reward-free pretraining method for autonomous driving using latent disagreement as intrinsic motivation, leading to improved zero-shot transfer and few-shot adaptation in unseen environments.

Contribution

InDRiVE is the first to leverage latent ensemble disagreement as an intrinsic motivation for reward-free pretraining in autonomous driving, enhancing transfer robustness.

Findings

Improved zero-shot robustness in unseen towns and routes.

Enhanced few-shot collision avoidance under town shift.

Disagreement-based pretraining outperforms reward-based methods.

Abstract

Model-based reinforcement learning (MBRL) can reduce interaction cost for autonomous driving by learning a predictive world model, but it typically still depends on task-specific rewards that are difficult to design and often brittle under distribution shift. This paper presents InDRiVE, a DreamerV3-style MBRL agent that performs reward-free pretraining in CARLA using only intrinsic motivation derived from latent ensemble disagreement. Disagreement acts as a proxy for epistemic uncertainty and drives the agent toward under-explored driving situations, while an imagination-based actor-critic learns a planner-free exploration policy directly from the learned world model. After intrinsic pretraining, we evaluate zero-shot transfer by freezing all parameters and deploying the pretrained exploration policy in unseen towns and routes. We then study few-shot adaptation by training a task policy with limited extrinsic feedback for downstream objectives (lane following and collision avoidance). Experiments in CARLA across towns, routes, and traffic densities show that disagreement-based pretraining yields stronger zero-shot robustness and robust few-shot collision avoidance under town shift and matched interaction budgets, supporting the use of intrinsic disagreement as a practical reward-free pretraining signal for reusable driving world models.