Hyundai Motor Group's Motional Leads the Culture of Autonomous Driving Data Sharing

Motional Releases Expanded Version of Autonomous Driving Data Library 'NuScene'
New Version Enhances LiDAR Data Collection and Incident Response
Strong Synergy Effect from Open-Source Autonomous Driving Data
Fostering a Global Autonomous Driving Ecosystem Data Sharing Culture

[Asia Economy Reporter Suyeon Woo] Motional, the autonomous driving joint venture (JV) of Hyundai Motor Group, recently released an expanded version of the data library 'nuScenes' that was open-sourced last year. Hyundai Motor Group plans to accelerate the commercialization of autonomous driving by advancing the data library, which will serve as a stepping stone for full autonomous driving, while also leading the establishment of a global autonomous driving ecosystem.

According to the automotive industry on the 8th, Motional recently unveiled a new version of nuScenes that includes 'nuScenes-lidarseg,' which enhances data collection using LiDAR, and 'nuImages,' a dataset specialized for handling unexpected situations.

Aptiv, the predecessor of Motional, had open-sourced the autonomous driving data library nuScenes in March last year to make it available to everyone. NuScenes contains over 1,000 driving scenes set in Boston, USA, and Singapore, along with 1.4 million image data points, radar signal analyses, and 390,000 LiDAR signal analyses.

At the time, Aptiv open-sourced the data library to assist startups that find it difficult to secure data independently in their technology development. One of the biggest challenges for autonomous driving startups in the early stages is the vast amount of data collection required. It takes considerable cost and time to equip expensive devices such as cameras, radar, and LiDAR and to gather data.

Aptiv, which was the first to open-source the collected data, established a 'data sharing culture' within the autonomous driving industry. Since the release of nuScenes last year, more than 8,000 researchers have used nuScenes for technology development, and over 250 papers have cited the data in their publications.

The open-source policy also benefited Aptiv. It was able to collect results obtained by various developers using nuScenes in diverse environments. Aptiv’s information-sharing policy has spread throughout the global autonomous driving industry, leading to the release of more than 10 data libraries, including those by Google Waymo.

The newly expanded nuScenes-lidarseg by Motional strengthens LiDAR environment optimization functions in the existing dataset. Until now, Motional’s technology had been evaluated as lacking in terms of data collection using LiDAR within the industry. However, through data sharing last year, they were able to quickly introduce a new version that reinforces the LiDAR segment.

The industry views the release of the new version just one year and six months after Motional first open-sourced nuScenes as a demonstration of the synergy effect of open source. Last month, Motional officially announced its company name and advanced the commercialization timeline for Level 4 autonomous driving from 2024 to 2022.

According to Navigant Research, Motional ranks joint 5th in autonomous driving technology strategy within the global autonomous driving industry. The top group includes Google Waymo, Ford, GM Cruise, and China’s Baidu, followed closely by Motional and Intel-Mobileye. This ranking places Motional ahead of latecomers such as Yandex, Volkswagen, Toyota, BMW, and Tesla.

Shin Yoon-chul, a researcher at Hi Investment & Securities, said, "Motional’s nuScenes has achieved rapid advancement recently because it was the first in the industry to open its data library. With the release of this expanded version, Motional has secured a favorable position to supply autonomous driving platforms to global original equipment manufacturers (OEMs)."