Competition among automakers, mobility platforms and software developers is intensifying ahead of broader adoption of autonomous driving. As vehicle platforms and AI models converge on common standards, companies are seeking advantage by accumulating real-world driving data to feed learning, verification and deployment. Hyundai Motor Group is pursuing a two-track approach using Nvidia’s platform while developing its own driving model. Kakao Mobility and RideFlux are building maps and operations and collecting difficult scenario data through commercial runs.

Kakao Mobility is shifting its autonomous driving development focus from collecting large volumes of driving data to selecting data from difficult situations, it said. Based on the data it has secured, the company plans to integrate perception and decision-making into a single end-to-end model within this year. An executive said the latest AI competition hinges on efficiently picking out hard data. The firm is classifying driving situations into five levels and prioritising rare edge-case data from higher levels.