Kakao Mobility is simultaneously pursuing the development of proprietary autonomous driving technology and the operation of a service platform in preparation for Level 4 driverless commercialization. By the end of this year, it plans to apply E2E (End-to-End) based driving technology to late-night autonomous vehicles in Gangnam and to build an integrated control system that manages autonomous vehicles from a single control center.

Since 2024, Kakao Mobility has been providing the Seoul Autonomous Driving Service through the Kakao T app. Autonomous driving technology companies can utilize Kakao T without separate customer or driver apps. The utilization rate of these vehicles is 84.3%. Currently, 97% of users of the Gangnam late-night Seoul Autonomous Vehicles in operation have used Kakao T without installing a separate app. The cumulative user rating stands at 4.97 out of 5.

Kakao Mobility believes that once Level 4 autonomous driving is commercialized, an operating system will be required that manages not only the vehicle’s driving technology but also ride-hailing and dispatching, control, and on-site response. Level 4 refers to a stage where the vehicle drives itself under specific conditions without driver intervention. Management is also needed for situations that vehicle AI alone finds difficult to handle, such as severe weather, signal failures, and traffic control.

Accordingly, Kakao Mobility is establishing a framework for real-time integrated monitoring of services and technology, 24-hour customer center (CS) support, on-site dispatch, and remote support from the control center. It also plans to implement an “Autonomous Integrated Control System (AICS)” that will centrally manage autonomous vehicles from a single control center.

In-house technology development is also under way. Since March, Kakao Mobility has been operating Seoul Autonomous Vehicles during late-night hours in Gangnam to collect data on unexpected urban situations such as two-wheeled vehicles, illegal parking and stopping, and jaywalking. Since June, it has also been concurrently training using daytime driving data.

The E2E (End-to-End) based driving technology to be applied at the end of this year uses a single AI model to process the entire sequence in which the vehicle recognizes its surroundings and determines its driving direction, with a rule-based Safety Evaluator conducting the final verification.

Lim In-ho, leader of the AI Driving Part within the Autonomous Driving Development Team at Kakao Mobility, said, “Because we have built a full stack of technologies over the past nine years, from sensors to vehicle control, we have been able to rapidly introduce E2E autonomous driving technology.”

Kim Min-sun, leader of the Autonomous Driving Business Team at Kakao Mobility, said, “For Level 4 driverless commercialization, the key question ultimately is ‘who will monitor and respond around the clock, and how,’” adding, “In the Level 4 era, core competitiveness will lie not only in driving AI, but also in integrated control and on-site response capabilities.”

Popular News