Citation: Yanase, R.; Hirano, D.;
Aldibaja, M.; Yoneda, K.; Suganuma,
N. LiDAR- and Radar-Based Robust
Vehicle Localization with Confidence
Estimation of Matching Results.
Sensors 2022, 22, 3545. https://
doi.org/10.3390/s22093545
Academic Editor: Andrzej Stateczny
Received: 7 April 2022
Accepted: 3 May 2022
Published: 6 May 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
LiDAR- and Radar-Based Robust Vehicle Localization with
Confidence Estimation of Matching Results
Ryo Yanase
1,
*
,†
, Daichi Hirano
2,†
, Mohammad Aldibaja
1
, Keisuke Yoneda
3
and Naoki Suganuma
1
1
Advanced Mobility Research Institute, Kanazawa University, Kakuma-Machi,
Kanazawa 920-1192, Ishikawa, Japan; amroaldibaja@staff.kanazawa-u.ac.jp (M.A.);
suganuma@staff.kanazawa-u.ac.jp (N.S.)
2
Graduate School of Natural Science and Technology, Kanazawa University, Kakuma-Machi,
Kanazawa 920-1192, Ishikawa, Japan; d_hirano@stu.kanazawa-u.ac.jp
3
Institute for Frontier Science Initiative, Kanazawa University, Kakuma-Machi,
Kanazawa 920-1192, Ishikawa, Japan; k.yoneda@staff.kanazawa-u.ac.jp
* Correspondence: ryanase@staff.kanazawa-u.ac.jp; Tel.: +81-76-234-4728
† These authors contributed equally to this work.
Abstract:
Localization is an important technology for autonomous driving. Map-matching using
road surface pattern features gives accurate position estimation and has been used in autonomous
driving tests on public roads. To provide highly safe autonomous driving, localization technology
that is not affected by the environment is required. In particular, in snowy environments, the features
of the road surface pattern may not be used for matching because the road surface is hidden. In such
cases, it is necessary to construct a robust system by rejecting the matching results or making up
for them with other sensors. On the other hand, millimeter-wave radar-based localization methods
are not as accurate as LiDAR-based methods due to their ranging accuracy, but it has successfully
achieved autonomous driving in snowy environments. Therefore, this paper proposes a localization
method that combines LiDAR and millimeter-wave radar. We constructed a system that emphasizes
LiDAR-based matching results during normal conditions when the road surface pattern is visible and
emphasizes radar matching results when the road surface is not visible due to snow cover or other
factors. This method achieves an accuracy that allows autonomous driving to continue regardless of
normal or snowy conditions and more robust position estimation.
Keywords: localization; sensor fusion; autonomous driving
1. Introduction
In autonomous driving, localization is important for path planning, decision-making,
operation, etc. Localization can be divided into two main types of methods. The first
is satellite positioning using the Global Navigation Satellite System (GNSS), which is
not available in places where radio waves do not reach, such as tunnels and mountain
areas, and in high buildings, where multipath can decrease the estimation accuracy [
1
]. The
second is map matching which estimates where the vehicle is located on a map by matching
sensor data with a map related in advance. Commonly used features in autonomous driving
include road surface patterns such as white lines and three-dimensional structures such as
poles and buildings. In many cases, autonomous driving is achieved by combining satellite
positioning and map matching, where satellite positioning is used to initialize and roughly
determine location, and map matching is used for more precise position estimation. Thus,
the challenge for safe autonomous driving is to increase the accuracy and robustness of
map matching.
The 2007 Urban Challenge was the first demonstration of autonomous driving in an
urban environment, and various universities and companies participated [
2
–
4
]. In this
project, Levinson et al. developed a road surface pattern-based map matching technique
Sensors 2022, 22, 3545. https://doi.org/10.3390/s22093545 https://www.mdpi.com/journal/sensors
