Max Robotics — Global Service Robot Directory + US Certification

About

Zhengtong Xu is a rising star in robotics whose research sits at the intersection of tactile sensing, learning-based control, and safe human-robot collaboration. His work centers on endowing robots with the ability to perceive, grasp, and manipulate objects with human-like dexterity and safety. Xu’s most impactful contribution is the VisTac system (28 citations), a unified multimodal sensing finger that seamlessly integrates tactile and proximity perception, enabling robots to sense objects both before and during contact. Building on this, he developed LeTac-MPC (15 citations), a learning-based model predictive control framework that uses tactile feedback to achieve robust, reactive grasping across diverse objects. His LeTO method introduces a novel approach to constrained visuomotor policy learning by embedding differentiable trajectory optimization directly into neural networks. Addressing the critical challenge of safety, Xu’s work on risk-tunable control barrier functions provides a principled framework for safe human-robot collaboration under uncertainty. Most recently, his UniT system demonstrates remarkable data efficiency, learning generalizable tactile representations from a single object. With publications appearing in top venues and a clear trajectory of innovation, Xu is establishing himself as a leading voice in tactile robotics and safe autonomous manipulation.