Kana Kim

Modular full-stack built for public-road driving in Songdo

Built a modular urban autonomous driving full-stack (perception, decision, planning, control, HD map, safety) across 100+ modules, validated through MORAI SIL and then on two public-road sections in Songdo.

Three-year lead, through two-vehicle public-road validation

Designed SAE J2735-based V2X messaging and the CLM (cooperative lane merging) and ETrA (emergency trajectory alignment) algorithms, then validated six cooperative scenarios with two vehicles (Ioniq5, Avante CN7) across K-City, KIAPI, Yeongjong, and Songdo.

Comparative profiling that quantifies the information gain of cooperative driving

To quantify why and how much cooperative driving helps, designed a 3-axis framework (R: input reliability, O: perception state-estimation quality, P: driving performance) with an OST/ΔOST metric, validated cooperative vs non-cooperative (WC vs WOC) across SIM, SIM-NPC, and real-vehicle (FIELD).

Two-hour driverless run in the rain on a high-speed track (Minister’s Award)

Implemented lane-specific driving strategies, curvature-based speed control, and a 50Hz Adaptive PID / Pure Pursuit controller to achieve a 2-hour driverless run at 100km/h (~129.2km) in the rain on the KIAPI high-speed circuit, finishing runner-up.

Automatically scoring and cleaning data quality without human review

After leading the collection and verification of a 930K-frame autonomous driving dataset, built ReliableDataset — a pipeline that removes noisy labels via model ensembles and whitens the data distribution to automatically score and prune the value of training data.

Level 3 system built through to a national operation permit

Implemented the full Sensing→Perception→Planning→Control→Safety→UI architecture, dual-CAN vehicle control, and AEB·BSD·LDW·TOR safety functions, passing four official tests at K-City and KIAPI and obtaining a MOLIT autonomous-driving temporary operation permit.