Identification of Road Surface Condition on Undeveloped Roads : - Aiming for Remote Car Driving

Autonomous vehicle developments have been actively pursued as a means to reduce the traffic accident numbers and alleviate environmental issues. However, even though some previous studies have attempted to identify issues related to undeveloped road environments, current autonomous driving models re...

Full description

Saved in:

Bibliographic Details
Published in:	2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) pp. 777 - 781
Main Authors:	Higashimoto, Kota, Fukushima, Hiroki, Kamitani, Kazumasa, Chujo, Naoya
Format:	Conference Proceeding
Language:	English
Published:	IEEE 12-10-2021
Subjects:	Autonomous driving Distance measurement Laser radar Point cloud compression Prototypes Real-time systems Remote car driving Road surface identification Roads Undeveloped roads Virtual environments
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Autonomous vehicle developments have been actively pursued as a means to reduce the traffic accident numbers and alleviate environmental issues. However, even though some previous studies have attempted to identify issues related to undeveloped road environments, current autonomous driving models require point cloud data from light detection and ranging (LiDAR) systems to be obtained in advance, which makes it difficult to identify undeveloped road environments in real time.With this point in mind, we propose an autonomous driving system with remote driving support that is designed for use on undeveloped roads in residential areas. More specifically, this system is intended to provide remote driving support to autonomous vehicles operating on undeveloped roads at speeds below 10 km/h.In our method, in which road environment identification is performed in real time, we identify developed and gravel surfaces as road surface types. We also identify potholes, humps, and street gutters as road surface obstacles.To evaluate our prototype system, experiments were conducted to determine if our method could identify road surfaces types in a real environment. Separately, road surface obstacle identification experiments were conducted in a virtual environment. The results of these experiments show that both surface types and obstacles could be identified with high probability. We also found that these identification processes could be run in real time. For street gutters, point cloud data in a real environment were obtained and discussed.
DOI:	10.1109/GCCE53005.2021.9621967