XYZ Joins Robotis in a Low-Cost Open-Source Humanoid Project
AI robotics firm XYZ will collect robot-learning data for a 4 billion won project, led by Robotis, to build an affordable open-source full-body humanoid platform.
SUMO· the number puzzle
Dialo· learn languages by speaking
ediviewer· read EDI files like documents
giftdetect· the right gift in minutesXYZ, a South Korean AI robotics company, said on the 5th that it has joined a 4 billion won (about) project to develop an open-source, full-body humanoid platform as a co-research institution. The project is led by Robotis, and XYZ will be responsible for securing the data used to train the robot.
The work is part of the Ministry of Trade, Industry and Resources' robot industry technology development program, with the Korea Evaluation Institute of Industrial Technology acting as the managing agency. Kwangwoon University and POSTECH are also co-research institutions. The official project title is development of an open-source, low-cost humanoid platform for walking reinforcement-learning research and manipulation imitation-learning data.
Hardware, software and data in one package
The participants plan to build an affordable bipedal humanoid using domestically made core components, together with open-source hardware and software that anyone can use. A full-body platform means walking, whole-body motion and object manipulation with arms and hands can be studied on a single robot. Behavior data and a simulation environment will be provided too, so researchers can validate and share results in a common development setting.
XYZ's data work has two tracks. The first is walking reinforcement learning, in which a robot learns motions through repeated trial and error; XYZ will handle moving gait learned in simulation onto a real robot. The second is manipulation imitation learning, where a person demonstrates tasks by teleoperating the robot and the robot learns from them. XYZ will gather those demonstrations.
Data will be collected in settings that reflect real use, such as stores and offices, while varying the objects, surroundings and demonstrators. After quality checks and standardization, it will be posted to the project's public data hub for use by other researchers and developers.
Training courses
XYZ will also design step-by-step courses covering robot control and simulation, reinforcement learning, imitation-learning data collection and deployment on real robots. Data produced during training will feed back into the research. Textbooks, exercises, simulation environments and operating manuals will be bundled and distributed to universities and professional training institutions.
"To widen the use of humanoids, it is important to connect hardware and control technology with learning data from varied environments," said CEO Hwang Seong-jae, adding that the company aims to build a data foundation spanning walking and manipulation and help the open-source humanoid ecosystem grow.
Apps worth a look
SUMO: the number puzzleTap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.
Dialo: learn languages by speakingTalk to an AI conversation partner; every sentence comes back corrected and explained in your language. 9 languages, A1 to C2.
ediviewer: read EDI files like documentsOpen EDIFACT and X12 messages, see every segment explained, convert to JSON/XML, export PDF and Excel. Runs locally, no upload.
giftdetect: the right gift in minutesPick who it is for, the occasion and their personality. The search engine reads product descriptions from local shops and ranks what fits. Free, no sign-up.
