← Company intelligence

Physical AI software · early

Deep-fusion AI

Corporate marketperceptionCurated record

Deep-fusion AI builds RAPA (Real-time Attention-based Pillar Architecture), a radar-centered sensor-fusion perception and 4D-imaging-radar SLAM software engine for autonomous vehicles and robots. RAPA-R fuses multiple 4D imaging radars for 360-degree omnidirectional perception, obstacle detection/classification and real-time mapping/localization, running on edge devices. It targets autonomous driving, logistics robots, and unmanned ground/surface vehicles (UGV/USV) rather than shipping any sensor hardware. Radar-camera (RAPA-RC) and radar-LiDAR (RAPA-RL) fusion variants are planned for 2027.

HeadquartersIncheon, KR
Founded2023
Corporate fundingNot disclosed
TeamNot listed
Deployments0
Open roles0
01

Company overview

Identity and operating footprint

Company typePhysical AI software
Market segmentCorporate
StageSeed (amount uncitable — aggregator-only)
FoundersSung Hun Yu (CEO)
Regions servedSouth Korea, Europe
Countries deployedNot listed
Service footprintNot listed
Last reviewed2026-07-16
CA

Funding history

Publicly announced rounds

Investors on recordTIPS (Tech Incubator Program for Startup)JNP GlobalTheVenturesY&Archer
TL

Company timeline

Milestones and announcements

2026-01-02DeepFusion AI to showcase RAPA-R multi-radar perception, proven in defense USV/UGV projects, at CES 2026AVING News
2025-11-19RAPA (Real-time Attention-based Pillar Architecture for 4D Radar Perception) named CES 2026 Innovation Award honoreeCES / Consumer Technology Association
2025-01-20Deep-fusion AI showcases 4D imaging radar deep-learning perception and real-time robot spatial mapping at CES 2025AVING News
SC

Source ledger

5 unique public sources