Centaur: Robust End-to-End Autonomous Driving with Test-Time Training

2025-03-14

AI Summary

To address low decision reliability and excessive behavioral conservatism in end-to-end autonomous driving deployment, this paper abandons hand-crafted rules and manually designed cost functions, proposing a test-time training (TTT)-based online adaptation method. Methodologically, it introduces cluster entropy as a novel uncertainty metric—first applied in this context—and dynamically fine-tunes the planner via unsupervised, single-step gradient updates that minimize this entropy. The approach achieves state-of-the-art performance on the navtest benchmark, ranking first and significantly improving critical safety metrics such as time-to-collision (TTC). Furthermore, the paper introduces navsafe, a new robustness evaluation benchmark that systematically uncovers previously undetected failure modes of existing end-to-end models, thereby advancing standardization in autonomous driving robustness assessment.

Technology Category

• Natural Language Processing: Safety and Robustness
• Planning, Routing, and Scheduling: Planning under Uncertainty
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

How can we rely on an end-to-end autonomous vehicle's complex decision-making system during deployment? One common solution is to have a ``fallback layer'' that checks the planned trajectory for rule violations and replaces it with a pre-defined safe action if necessary. Another approach involves adjusting the planner's decisions to minimize a pre-defined ``cost function'' using additional system predictions such as road layouts and detected obstacles. However, these pre-programmed rules or cost functions cannot learn and improve with new training data, often resulting in overly conservative behaviors. In this work, we propose Centaur (Cluster Entropy for Test-time trAining using Uncertainty) which updates a planner's behavior via test-time training, without relying on hand-engineered rules or cost functions. Instead, we measure and minimize the uncertainty in the planner's decisions. For this, we develop a novel uncertainty measure, called Cluster Entropy, which is simple, interpretable, and compatible with state-of-the-art planning algorithms. Using data collected at prior test-time time-steps, we perform an update to the model's parameters using a gradient that minimizes the Cluster Entropy. With only this sole gradient update prior to inference, Centaur exhibits significant improvements, ranking first on the navtest leaderboard with notable gains in safety-critical metrics such as time to collision. To provide detailed insights on a per-scenario basis, we also introduce navsafe, a challenging new benchmark, which highlights previously undiscovered failure modes of driving models.

Problem

• Improves autonomous vehicle decision-making via test-time training.
• Minimizes decision uncertainty using Cluster Entropy metric.
• Enhances safety metrics without pre-programmed rules or cost functions.

Innovation

• Test-time training updates planner behavior
• Cluster Entropy minimizes decision uncertainty
• Gradient update improves safety-critical metrics