Does Knowledge About Perceptual Uncertainty Help an Agent in Automated Driving?

2025-02-17

AI Summary

This work addresses the problem of insufficient integration of perception uncertainty into decision-making for autonomous driving. We propose an uncertainty-aware reinforcement learning framework that explicitly encodes and injects perception uncertainty estimates into the observation space of an end-to-end driving policy—marking the first such approach. Our method combines controllable synthetic perturbations with large-scale simulation experiments in CARLA and Highway-env to jointly achieve perception calibration and goal-directed control. Under zero-collision constraints, our agent reduces average passage time by 19% compared to baselines, demonstrating superior risk identification and adaptive aggressiveness modulation. The core contribution is a novel, interpretable mapping mechanism that bridges perception uncertainty to behavioral decisions, establishing a principled paradigm for safety–efficiency trade-offs in autonomous driving.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Planning, Routing, and Scheduling: Planning under Uncertainty
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Agents in real-world scenarios like automated driving deal with uncertainty in their environment, in particular due to perceptual uncertainty. Although, reinforcement learning is dedicated to autonomous decision-making under uncertainty these algorithms are typically not informed about the uncertainty currently contained in their environment. On the other hand, uncertainty estimation for perception itself is typically directly evaluated in the perception domain, e.g., in terms of false positive detection rates or calibration errors based on camera images. Its use for deciding on goal-oriented actions remains largely unstudied. In this paper, we investigate how an agent's behavior is influenced by an uncertain perception and how this behavior changes if information about this uncertainty is available. Therefore, we consider a proxy task, where the agent is rewarded for driving a route as fast as possible without colliding with other road users. For controlled experiments, we introduce uncertainty in the observation space by perturbing the perception of the given agent while informing the latter. Our experiments show that an unreliable observation space modeled by a perturbed perception leads to a defensive driving behavior of the agent. Furthermore, when adding the information about the current uncertainty directly to the observation space, the agent adapts to the specific situation and in general accomplishes its task faster while, at the same time, accounting for risks.

Problem

• Impact of perceptual uncertainty on agent behavior
• Influence of uncertainty information on decision-making
• Enhancing automated driving with uncertainty awareness

Innovation

• Integrates uncertainty into reinforcement learning.
• Uses perturbed perception for controlled experiments.
• Enhances decision-making with uncertainty information.