Neural Semantic Map-Learning for Autonomous Vehicles

TL;DR

This paper introduces a neural map-learning system for autonomous vehicles that fuses crowd-sourced local submaps into a coherent 3D environment model, improving map accuracy and scalability.

Contribution

It presents a novel neural signed distance field approach that aligns and merges local submaps from multiple vehicles, scalable with sparse feature grids and uncertainty modeling.

Findings

Enhanced pose alignment and reconstruction accuracy.

Effective multi-session mapping capabilities.

Scalable to large areas with sparse features.

Abstract

Autonomous vehicles demand detailed maps to maneuver reliably through traffic, which need to be kept up-to-date to ensure a safe operation. A promising way to adapt the maps to the ever-changing road-network is to use crowd-sourced data from a fleet of vehicles. In this work, we present a mapping system that fuses local submaps gathered from a fleet of vehicles at a central instance to produce a coherent map of the road environment including drivable area, lane markings, poles, obstacles and more as a 3D mesh. Each vehicle contributes locally reconstructed submaps as lightweight meshes, making our method applicable to a wide range of reconstruction methods and sensor modalities. Our method jointly aligns and merges the noisy and incomplete local submaps using a scene-specific Neural Signed Distance Field, which is supervised using the submap meshes to predict a fused environment representation. We leverage memory-efficient sparse feature-grids to scale to large areas and introduce a confidence score to model uncertainty in scene reconstruction. Our approach is evaluated on two datasets with different local mapping methods, showing improved pose alignment and reconstruction over existing methods. Additionally, we demonstrate the benefit of multi-session mapping and examine the required amount of data to enable high-fidelity map learning for autonomous vehicles.