Improved Radar Localization on Lidar Maps Using Shared Embedding

TL;DR

This paper introduces a novel heterogeneous localization framework that uses shared embedding spaces created by deep neural networks to improve radar-based localization and pose tracking on lidar maps, demonstrating superior performance over existing methods.

Contribution

The paper proposes a new deep learning-based framework that creates shared embeddings for radar and lidar data, enhancing localization accuracy and reducing neural network complexity.

Findings

Outperforms Scan Context and RaLL in experiments

Effective radar localization on lidar maps demonstrated on datasets

Reduced neural network complexity in pose tracking pipeline

Abstract

We present a heterogeneous localization framework for solving radar global localization and pose tracking on pre-built lidar maps. To bridge the gap of sensing modalities, deep neural networks are constructed to create shared embedding space for radar scans and lidar maps. Herein learned feature embeddings are supportive for similarity measurement, thus improving map retrieval and data matching respectively. In RobotCar and MulRan datasets, we demonstrate the effectiveness of the proposed framework with the comparison to Scan Context and RaLL. In addition, the proposed pose tracking pipeline is with less neural networks compared to the original RaLL.