Max Robotics — Global Service Robot Directory + US Certification

About

David B. Grimes is a pioneering robotics and machine learning researcher whose work centers on imitation learning, probabilistic inference, and humanoid robot control. His research has made substantial contributions to one of robotics' most ambitious challenges: enabling machines to learn complex behaviors by observing humans, rather than through exhaustive manual programming. Grimes's most influential work applies Bayesian and nonparametric probabilistic frameworks to humanoid motion learning, most notably demonstrated in his 2006 paper on dynamic imitation using Bayesian networks (98 citations), which showed how rich human motion capture data could guide whole-body robot movement under uncertainty. Complementing this, his probabilistic model of gaze imitation and shared attention (81 citations) addressed the socially critical capacity for robots to follow and mirror human eye gaze — a foundational element of natural human-robot interaction. His 2007 contribution demonstrating imitation-based humanoid walking (44 citations) broke new ground as one of the first systems to achieve stable bipedal locomotion learned directly from human demonstration. Across more than a dozen publications accumulating over 380 citations, Grimes consistently advanced the idea that uncertainty-aware, probabilistic models offer the most principled path toward robots that genuinely learn from human teachers — a vision that continues to shape modern imitation learning research.