Max Robotics — Global Service Robot Directory + US Certification

About

Mitchell Hebert is a robotics researcher whose work bridges the critical gap between autonomous perception and safe human-robot collaboration. His primary research areas include semantic perception for manipulation, human-robot teaming, and competency-aware autonomy. Hebert's most impactful contribution is **SegICP** (151 citations), a pioneering framework that integrates deep semantic segmentation with pose estimation, enabling robots to rapidly and reliably perceive objects in cluttered, realistic environments—a breakthrough directly addressing bottlenecks identified in robotic manipulation competitions. He extended this work with **SegICP-DSR**, achieving millimeter-level pose accuracy for dense semantic scene reconstruction. More recently, Hebert has focused on the human side of autonomy, developing methods for **generalizing competency self-assessment** in autonomous vehicles using deep reinforcement learning, and critically examining **human non-compliance with robot spatial ownership** communicated via augmented reality—work with direct implications for safety in human-robot teams. His research on collaborative planning and negotiation further targets high-risk environments like space operations. Through this trajectory, Hebert demonstrates a rare ability to advance both the perceptual capabilities of robots and the foundational trust and safety mechanisms essential for their real-world deployment alongside humans.