Delving into Mapping Uncertainty for Mapless Trajectory Prediction

IROS 2025 Submission

Zongzheng Zhang^1,2* · Xuchong Qiu^2* · Boran Zhang¹ · Guantian Zheng¹ · Xunjiang Gu⁴
Guoxuan Chi¹ · Huan-ang Gao¹ · Leichen Wang² · Ziming Liu¹ · Xinrun Li²
Igor Gilitschenski⁴ · Hongyang Li⁵ · Hang Zhao³ · Hao Zhao¹

¹ Institute for AI Industry Research (AIR), Tsinghua University ² Bosch Corporate Research
³ Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University
⁴ University of Toronto ⁵ The University of Hong Kong

^* Equal contribution

Code

Dataset

Motivation

(a) Motivation of the Scenario Classifier

(b) Motivation of the Covariance-Based Uncertainty

Left: Mapless trajectory prediction baseline (MapTRv2 + HiVT), Middle: Previous uncertainty integration method, Right: Ours.
Pink indicates predicted trajectories, red shows the ground truth.
In (a), the comparison highlights that predictions enhanced with uncertainty occasionally underperform the baseline, underscoring the necessity of a gating mechanism to selectively incorporate uncertainty.
In (b), our Covariance-based Map Uncertainty captures road curvature with good precision that the predicted trajectories align well with the ground truth, rendering the true trajectory invisible.

Abstract

Recent advances in autonomous driving are moving towards mapless approaches, where High-Definition (HD) maps are generated online directly from sensor data, reducing the need for expensive labeling and maintenance. However, the reliability of these online-generated maps remains uncertain. While incorporating map uncertainty into downstream tra- jectory prediction tasks has shown potential for performance improvements, current strategies provide limited insights into the specific scenarios where this uncertainty is beneficial. In this work, we first analyze the driving scenarios in which mapping uncertainty has the greatest positive impact on trajectory prediction and identify a critical, previously overlooked factor: the agent's kinematic state. Building on these insights, we propose a novel Proprioceptive Scenario Gating that adaptively integrates map uncertainty into trajectory prediction based on forecasts of the ego vehicle's future kinematics. This lightweight, self-supervised approach enhances the synergy between online mapping and trajectory prediction, providing interpretability around where uncertainty is advantageous and outperforming previous integration methods. Additionally, we introduce a Covariance-based Map Uncertainty approach that better aligns with map geometry, further improving trajectory prediction. Extensive ablation studies confirm the effectiveness of our approach, achieving up to 23.6% improvement in mapless tra- jectory prediction performance over the state-of-the-art method using the real-world nuScenes driving dataset.

Overview

We propose a lightweight, self-supervised approach that enhances the synergy between online mapping and trajectory prediction, providing interpretability on when and where map uncertainty is beneficial. Additionally, we introduce a covariance-based uncertainty modeling method that better aligns with road geometry. Extensive ablation studies show that our method outperforms previous integration strategies, achieving up to 23.6% improvement in mapless trajectory prediction on the nuScenes dataset.