UBTech launches Thinker-Sim embodied AI simulation platform spanning scene digitization and capability evaluation
UBTech released the Thinker-Sim embodied AI simulation platform on November 2, a one-stop technical system covering real-world scene digitization, large-scale synthetic simulation data and robot capability evaluation.
UBTech released the Thinker-Sim embodied AI simulation platform on November 2, a one-stop technical system covering real-world scene digitization, large-scale synthetic simulation data and robot capability evaluation. It uses UBTech's in-house Thinker-GS high-fidelity scene reconstruction algorithm and the full-pipeline SimReady automated asset generation to convert real scene capture and 3D reconstruction into simulation assets efficiently. Reconstructing from actual factory production lines restores plant layout, structure and visual detail, and foreground assets such as robots, production equipment and personnel can be loaded flexibly for training, data collection or algorithm evaluation. The platform can also generate digital twin environments with realistic visuals for commercial service scenarios, supporting robot training and deployment verification across settings, and combines image-text-driven batch 3D asset generation, low-poly topology optimization, visual material generation and automatic collider configuration to balance visual quality, physics fidelity and production efficiency. Built-in standardized task templates cover common training, data synthesis and evaluation scenarios so users can start simulations by selecting templates and setting parameters without building environments from scratch, with live frame streaming, task status monitoring and result feedback.