Max Robotics — Global Service Robot Directory + US Certification

About

Mathew Halm is a robotics researcher whose work tackles one of the field’s most persistent challenges: modeling and learning the physics of contact. His research focuses on the fundamental dynamics of frictional impact, stiction, and nearly discontinuous motion—phenomena that underpin legged locomotion and dexterous manipulation. Halm’s major contributions include pioneering implicit learning frameworks for contact dynamics, most notably through ContactNets (2020, 18 citations), which resolves the conflict between smooth neural network predictions and the discontinuous reality of impact and stiction. His 2021 study on fundamental challenges in deep learning for stiff contact dynamics (20 citations) provides critical empirical evidence that standard learning approaches fail for contact-rich tasks, guiding the community toward more robust methods. Halm has also advanced the theory of simultaneous frictional impacts, developing set-valued rigid-body dynamics (2021, 7 citations; 2024, 5 citations) that address the extreme sensitivity of collision outcomes to impact ordering—a problem that plagues robotic simulators and state estimation. His theoretical work on generalization bounds for implicit learning of nearly discontinuous functions (2021) offers rigorous foundations for these empirical successes. Through this combination of theoretical insight and practical methodology, Halm is shaping how robots learn to reliably interact with the physical world.