Risk-Aware World Model Predictive Control for Generalizable End-to-End Autonomous Driving

arXiv 26.2 2026 52.6 method, application

TLDR

Proposes RaWMPC, a risk-aware world model predictive control framework for end-to-end autonomous driving without expert demonstrations.

Reasoning

The paper addresses generalization in long-tail scenarios by using a world model for risk evaluation, which is a strength. However, the abstract lacks details on real-world validation and baseline comparisons, limiting assessment of practical impact.

Read-first score

Read-first score 52.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 29.

Recency 8%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=dataset,evaluation,experiment

Topical relevance 42%

41.4

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=dataset

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 473.

Keyword Scores

world model

world dynamics prediction

interactive world model

world simulator

generative world model

model-based reinforcement learning world model

video world model

Deep Analysis

Innovations

Risk-aware World Model Predictive Control (RaWMPC) framework that enables end-to-end autonomous driving without reliance on expert demonstrations.
Risk-aware interaction strategy that systematically exposes the world model to hazardous behaviors, making catastrophic outcomes predictable and avoidable.
Self-evaluation distillation method to distill risk-avoidance capabilities from the world model into a generative action proposal network, eliminating the need for expert action supervision.

Methodology

RaWMPC uses a world model to predict the consequences of multiple candidate actions and selects low-risk actions through explicit risk evaluation. A risk-aware interaction strategy exposes the world model to hazardous behaviors during training to improve prediction of risky outcomes. At test time, a self-evaluation distillation method transfers risk-avoidance capabilities from the world model to a generative action proposal network, enabling low-risk action generation without expert demonstrations.

Key Results

RaWMPC outperforms state-of-the-art methods in both in-distribution and out-of-distribution scenarios, while providing superior decision interpretability.

Resources

Paper