Exploring Real World Map Change Generalization of Prior-Informed HD Map Prediction Models

TL;DR

This paper investigates how well prior-informed HD map prediction models trained on synthetic perturbations generalize to real-world map changes, revealing a significant sim2real gap that challenges autonomous vehicle deployment.

Contribution

It provides a large-scale experimental analysis of synthetic perturbations' effectiveness in real-world map change prediction, highlighting the gap and limitations of current models.

Findings

Synthetic perturbations have limited effectiveness in real-world generalization.

A substantial gap exists between synthetic training and real-world map changes.

Current models struggle to accurately predict real-world HD map updates.

Abstract

Building and maintaining High-Definition (HD) maps represents a large barrier to autonomous vehicle deployment. This, along with advances in modern online map detection models, has sparked renewed interest in the online mapping problem. However, effectively predicting online maps at a high enough quality to enable safe, driverless deployments remains a significant challenge. Recent work on these models proposes training robust online mapping systems using low quality map priors with synthetic perturbations in an attempt to simulate out-of-date HD map priors. In this paper, we investigate how models trained on these synthetically perturbed map priors generalize to performance on deployment-scale, real world map changes. We present a large-scale experimental study to determine which synthetic perturbations are most useful in generalizing to real world HD map changes, evaluated using multiple years of real-world autonomous driving data. We show there is still a substantial sim2real gap between synthetic prior perturbations and observed real-world changes, which limits the utility of current prior-informed HD map prediction models.