XYZ Unveils Real-time Robot Imitation Learning Platform 'TwinX'

XYZ, a company specializing in AI robot solutions, is receiving significant attention after unveiling 'TwinX,' its Unity-based remote teaching and imitation learning environment that will revolutionize the future of the robotics industry, on the 27th. This system was unveiled at a robotics industry innovation event held with Unity, and its significance is unique in that it goes beyond simple virtual simulation and has been directly applied to a real commercial store environment. XYZ has perfectly implemented the 'LoungeX 24h' store, where its service robot 'BaristaBot' operates, as a sophisticated digital twin environment capable of real-time teleoperation. This innovative method involves remotely teaching a robot by applying the actions of a robot operator in real-time to a virtualized robot in a real store, and then building a learning model based on the accumulated movement data through this process. This endeavor holds high value, both industrially and commercially, because robot intelligence and control were developed based on a commercial store where actual customer interaction occurs. The 'TwinX Remote Teaching/Imitation Learning System' has dramatically improved learning efficiency and precision compared to existing manual teaching methods, and boasts scalability for universal application to various types of robots in the future, including cleaning robots, delivery robots, dual-arm manipulation robots, and even humanoids. Jo In-sung, head of XYZ's Robot Intelligence Research Institute, emphasized, "This is an example demonstrating that real-data-based robot intelligence can provide a practical solution to the real-world problems of the service robot industry." XYZ has, until now, developed robot-user interaction models, multimodal interfaces, and digital twin-based learning systems as core technologies, and this technology unveiling is the first result of integrating all these elements into a single commercial store. Alongside this, XYZ announced its plan to develop an 'Interactive Robot Foundation Model for Learning Emotional Empathy,' which will evolve robots to understand and empathize with emotions and situations based on multimodal data (voice, vision, robot movements, etc.) collected from customer interactions, thereby opening new horizons for robot AI. This aligns with the recent global trend of Vision-Language-Action models such as RT-2, Mobile Aloha, and Open-VLA, which have garnered significant attention, and heralds an era of more natural human-robot interaction. Through its proprietary AI learning infrastructure and robot-user behavior data collection platform, XYZ is accelerating the development of robot intelligence and human-robot interaction models, and plans to further solidify its position in the global robot market. Through this event, XYZ unveiled its next-generation robot AI strategy and technology, and plans to expand its industrial collaborations and global expansion efforts.