A robot that reads the room: Ludi 0.2 learns gestures, intent and the pause mid-task

Most robots do exactly what they are told, even when the person telling them has changed their mind. Ludo Robotics, the robotics arm of South Korean game maker Krafton, is teaching its system to notice. Ludi 0.2, presented on 30 October at the Bay Area Robotics Symposium at UC Berkeley, reads human gestures, feedback and social cues, remembers the context when a task shifts midstream, and combines moving with handling objects. The team also opened a research lab in Palo Alto, California, to work alongside its Seoul office, and is building toward Ludi 1.0 — one model uniting seeing, talking and doing.

SDG 9 · Industry, innovation and infrastructure — Mechanism: tech innovation