Max Robotics — Global Service Robot Directory + US Certification

About

Mingxu Zhang is a rising researcher in embodied AI and robotic manipulation, whose work focuses on bridging large language models with physical robot control. His most impactful contribution, the highly cited "ManipLLM" (2024, 65 citations), introduces an embodied multimodal large language model that tackles a core challenge in robotics: enabling object-centric manipulation that generalizes beyond simulator-trained categories. By using LLMs to predict precise contact points and end-effector directions, Zhang’s approach moves away from rigid, category-specific training toward more flexible, real-world operation. He further advances this line of research with "CrayonRobo" (2025), which addresses ambiguity in task specification by combining language, goal images, and videos into a unified prompt-driven vision-language-action model. Complementing these works, his adaptive tracking control research (2024) provides a mathematical foundation for handling kinematic and dynamic uncertainties in robotic manipulators using zeroing-dynamics methods. Collectively, Zhang’s work demonstrates a clear trajectory from theoretical control to practical, generalizable manipulation systems, positioning him as a key contributor to the next generation of intelligent, adaptable robots.