Physical AI

POSCO DX is realizing autonomy and unmanned operation for physical facilities in heavy industries by integrating OT, IT, and AI technologies. In particular, it is pursuing technological advancement by establishing a Physical AI development system that utilizes physical development methodologies, such as virtual environment simulations.

Spatial Intelligence

Realizing unmanned manufacturing sites by implementing Physical AI based on self-developed deep learning architecture, utilizing field-specific data collected through key sensors installed in manufacturing sites such as LiDAR, Smart CCTV, and ToF.

Control convergence technology

Promoting unmanned facilities through the integration of robot and PLC control and vision technology

Reference
  • Hanjin Express Vision AI-enabled real-time parcel classification
  • Equipping steel unmanned mobile equipment using smart CCTVs with AI-enabled Foolproof
  • Unmaned through detection of 4-continuous cast slab meandering, etc.
  • 2,4-wire product inner/outer diameter shape measurement technology
  • Foolproof-based unmanned safety environment application technology
Sensor technology fusion

Securing multi-modal (Vision+X) technology to overcome limitations from a single fixed sensor

Reference
  • center point estimation technology for unmanned wire cranes
  • slurry yard mapping technology using LiDAR
  • Applying load estimation AI model to estimate loading & unloading times of Hanjin Express vehicles
  • Product shape recognition technology for unmanned backplate cranes