Baida leads KRW 11B AVM national project for retrofit driverless vehicle marshalling
The AMW Read
Updates the Robotics/Physical AI player map with a Korean radar-safety firm leading a KEIT detachable-AVM marshalling program; incremental industrial sub-segment signal, not a structural capital or geopolitics story.
Baida leads KRW 11B AVM national project for retrofit driverless vehicle marshalling
Baida (바이다), a radar-based mobility safety company, has been named lead research institution for a KEIT national R&D program titled development of multi-vehicle remote operation and safety technology based on detachable terminals. Total project cost is about 10.97 billion won (~$8.2M), with 9 billion won (~$6.7M) in government R&D support through December 2029. Seven Korean co-researchers join—including Saesol Tech, KAIST, KATECH, Autobiz Systems, DH Autoware, LG Electronics, and Renault Korea—alongside Czech partners Ostrava University of Technology, Intens, Skoda, and O2 Telefonica for joint industrial validation.
The target capability is Automated Vehicle Marshalling (AVM): moving finished vehicles without drivers inside factories and ports and parking them at designated spots. Instead of embedding full autonomy in every car or blanketing a yard with heavy infrastructure, the consortium will clip on a detachable terminal that leaves mass-production vehicle internals unchanged. The technical stack spans one-to-many remote supervision over V2N links, AI prediction of segment-level communication quality, infrastructure sensors to cover vehicle blind spots, a vision-language model that reads hazard and work context from video, and a zero-trust safety layer that verifies users, devices, and control messages and safely stops vehicles on anomalous access or link failure. That combination sits squarely in constrained-site physical AI—industrial yards, not open-road robotaxis—and aims to cut labor and safety risk in finished-vehicle logistics while lowering the capex barrier versus factory-fitted autonomy.
For builders and investors, the near-term commercial path is procurement and standards, not consumer AV narratives. Proving detachable AVM on stock OEMs, then feeding Korea–Czech field results into ISO, SAE, and UNECE proposals and OEM plant or port channels (Renault Korea, Skoda), is a clearer route to industrial revenue than general-purpose humanoid bets. Watch whether the detachable-terminal model actually clears latency, cybersecurity, and multi-vehicle control hurdles in live plant conditions by the 2029 deadline.