Doosan infracore Concept-X
Oct 14, 2021
1 min read
Updated: Jul 8, 2025
SNU-INRoL participated Doosan infracore's demonstration for futuristic construction site, named, "Concept-X". INRoL contributed the expert-emulating trajectory generation for autonomous excavators based on machine-learning and DMP (dynamic motion primitives) approaches. The results were also published and presented at IROS 2019.
Impressive application of smart construction technology! For clearer equipment and display photos, an online moiré removal tool can be useful.
Para quienes sienten curiosidad por estos temas, descubrir tu número personal puede convertirse en una actividad interesante. Además de conocer su interpretación simbólica, el proceso permite prestar atención a fechas y números que normalmente pasan desapercibidos. Puede ser una forma sencilla de dedicar unos minutos a la reflexión personal.
The Concept-X demonstration highlights how machine-learning and dynamic motion primitives can support expert-emulating trajectory generation for autonomous excavators. For a lighter change of pace afterward, Juegos de FNF offers browser-playable Friday Night Funkin’ mods and rhythm games for Spanish-speaking readers.
The dynamic motion primitives angle for excavator arms is a neat detail, applying motion primitives to heavy machinery rather than animation feels unexpected. I mostly see that technique in character rigging discussions. Between study breaks I've been messing around on fnf which is a rhythm game with rap battles and a very 90s visual style.
The DMP part caught my eye, using expert demonstrations to shape excavator motion instead of hand tuning every path is a practical way to handle messy terrain. Publishing it at IROS 2019 also means the approach got looked at by people outside the demo. I mess around with text to music on weekends, mainly feeding it hummed melodies to see what comes back, and it scratches a similar itch of getting a machine to imitate how a human does something.