The Center is dedicated to advancing AI's understanding of 3D space and physical laws. By building upon these capabilities, we aim to achieve breakthroughs in robotics, achieving agile motion control and fine manipulation skills for humanoid robots.

We will investigate the fundamental principles of multimodal generative models to develop general spatial intelligence foundation models. This effort seeks to extend the "pre-training plus fine-tuning" paradigm from language and image/video domains to 3D space.

Concurrently, we will explore data-driven robot learning theories, integrating multi-sensor perception such as vision and touch. Combined with motion control algorithms for humanoid robots and multi-fingered dexterous hands, we strive to enable robots with agile mobility and precise manipulation skills.

Leveraging these achievements, we will integrate with the strengths of the Greater Bay Area in mechatronics and manufacturing, fostering deep industry collaboration. Our goal is to promote the adoption and application of robotics technology across various sectors, including industrial, medical, elderly care, and service industries.