The Impact of Class Uncertainty Propagation in Perception-Based Motion Planning

2026-02-17

AI Summary

This work addresses the sensitivity of autonomous driving planning to poorly calibrated prediction uncertainty, which critically impacts safety and generalization. It presents the first systematic study of how perception uncertainty propagates and is calibrated through the prediction–planning closed loop, introducing two hierarchical uncertainty propagation pipelines. Their impact on motion planning performance is evaluated on the nuPlan benchmark using closed-loop simulation, co-designed pipeline architectures, and a dedicated calibration assessment methodology. The study reveals that explicit modeling of upstream uncertainty plays a pivotal role in enabling robust generalization to complex scenarios. Experimental results demonstrate that planning methods incorporating upstream uncertainty explicitly achieve significantly improved closed-loop generalization performance on nuPlan’s challenging scenarios.

Technology Category

• Planning, Routing, and Scheduling: Planning under Uncertainty
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations

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
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Autonomous vehicles (AVs) are being increasingly deployed in urban environments. In order to operate safely and reliably, AVs need to account for the inherent uncertainty associated with perceiving the world through sensor data and incorporate that into their decision-making process. Uncertainty-aware planners have recently been developed to account for upstream perception and prediction uncertainty. However, such planners may be sensitive to prediction uncertainty miscalibration, the magnitude of which has not yet been characterized. Towards this end, we perform a detailed analysis on the impact that perceptual uncertainty propagation and calibration has on perception-based motion planning. We do so by comparing two novel prediction-planning pipelines with varying levels of uncertainty propagation on the recently-released nuPlan planning benchmark. We study the impact of upstream uncertainty calibration using closed-loop evaluation on the nuPlan challenge scenarios. We find that the method incorporating upstream uncertainty propagation demonstrates superior generalization to complex closed-loop scenarios.

Problem

• class uncertainty
• uncertainty propagation
• perception-based motion planning
• uncertainty calibration
• autonomous vehicles

Innovation

• uncertainty propagation
• perception-based motion planning
• uncertainty calibration
• autonomous driving
• closed-loop evaluation