Laura Major, CEO of Motional, Hyundai Motor Group's Autonomous Driving joint venture, appeared on the group's official podcast channel "Hyundai Now" and again emphasized the direction of Autonomous Driving technology development and the commercialization strategy for robotaxis.
On the 8th, according to Hyundai Motor Group, Major said through Hyundai Now that Motional has taken a step forward by shifting the structure of its Autonomous Driving technology to center on a generative artificial intelligence (AI)-based large driving model (LDM), and stated accordingly. Motional is preparing to commercialize robotaxis by the end of this year.
Hyundai Now is a knowledge series themed on mobility technology, with a total of 13 episodes released so far. This is the 14th episode, and until now it had proceeded in a format where industry experts took part as panelists. This is the first time a senior Hyundai Motor Group executive has appeared directly to communicate. In this video, the hosts and Major held a Q&A via video conference.
Major emphasized, "In 2024, the decision to redesign Motional's Autonomous Driving system architecture around AI and shift to LDM was an important turning point in Autonomous Driving technology development." Changing the system architecture entails risk factors compared with previous work, she said, but successfully implementing the transition became an opportunity to take a step up in several areas, including driving quality, responsiveness, global scalability and operating costs.
Major highlighted passenger safety as the top principle in Autonomous Driving development. She said, "We apply an end-to-end (E2E) approach using the large driving model to the general driving situations that account for more than 90% of driving, and we introduced a long-validated safety guardrail approach for the 1% edge cases such as emergencies."
The guardrail approach refers to a safety shield regime applied first to prevent the Autonomous Driving system from making incorrect decisions. Major added, "The 1% edge cases that occur in exceptional situations are the key training data that improve Autonomous Driving performance." Saying that "fail fast" is Motional's creed, Major said, "The faster you find a problem, the faster you can fix it and the faster you can learn."
Motional currently operates research and development hubs in several regions of the United States, and is conducting intensive testing in two cities in particular, Las Vegas and Pittsburgh. Major said, "Las Vegas has modern characteristics such as a planned urban structure and wide roads, while Pittsburgh has an older urban environment with narrow, winding roads and complex intersections," adding, "By testing in contrasting environments, we can develop a general-purpose Autonomous Driving system applicable in any city."
Major cited data accumulated from real service experience as Motional's commercialization edge. Motional has run various pilot programs in cooperation with ride-hailing platforms such as Uber and Lyft, through which it has secured more than 130,000 Autonomous Driving service experiences.
She said, "In the process, we gained various insights, such as what information passengers want inside the vehicle, whether they use the vehicle display or their smartphone more, and how to implement functions such as route changes or mid-ride drop-offs in the system."
Major stressed that, in addition to developing Autonomous Driving technology, improving the passenger experience is also a key factor at the commercialization stage. She said, "We are carrying out continuous technology improvements with the goal of launching a driverless Autonomous Driving service by the end of this year and are focusing on advancing the overall passenger experience before commercialization."
Meanwhile, Major recently received the SDV Innovator Awards selected by the North American automotive magazine MotorTrend.