UBTech unveils thinker-sim, a comprehensive platform for embodied AI simulation and robot capability testing

UBTECH Technology has released Thinker-Sim, a unified technical system designed to cover digital representation of real-world scenarios, large-scale synthetic data generation, and rigorous robot capability evaluation for embodied AI. The platform leverages UBTECH's proprietary Thinker-GS high-fidelity scene reconstruction algorithm and end-to-end SimReady asset automation technology. This allows for an efficient conversion process, moving from real-world scene capture and 3D reconstruction directly into simulation assets. When reconstructing environments like factory production lines, the system can accurately reproduce spatial layouts, structural details, and visual nuances, while flexibly accommodating foreground assets such as robots, production machinery, and personnel based on specific training or testing needs. Furthermore, Thinker-Sim supports commercial applications by rapidly generating digital twin environments with realistic visual fidelity, enabling robot training and deployment verification across various settings. Key technical features include batch generation of 3D assets driven by text and images, low-poly topological optimization, visual material generation, and automatic collision body configuration, balancing visual quality with physical simulation and production efficiency. The system also features built-in standardized simulation task templates, allowing users to quickly initiate tasks through parameter configuration rather than building environments from scratch, complete with real-time stream monitoring and result feedback.