Max Robotics — Global Service Robot Directory + US Certification

About

Pedram Azad is a robotics researcher whose work sits at the intersection of humanoid robotics, computer vision, and machine learning, with particular expertise in imitation learning, human motion capture, and robot perception. His most influential contribution, "Imitation Learning of Dual-Arm Manipulation Tasks in Humanoid Robots" (2006, 158 citations), pioneered the use of Hidden Markov Models to enable robots to generalize from multiple human demonstrations — a foundational advance in programming robots by example. Complementing this, his work on unified motion representation frameworks and non-linear optimization techniques established robust pipelines for transferring human motor knowledge to humanoid platforms. Azad made significant strides in robot perception through his development of the Karlsruhe Humanoid Head (104 citations), a sophisticated sensory system integrated into the ARMAR robot series, and through stereo-based 6D object localization methods that enabled reliable autonomous grasping in real-world environments. His markerless human motion capture systems, using only stereo camera input, removed the need for intrusive tracking hardware — a practically meaningful achievement. Across his most-cited works, Azad has accumulated over 700 citations, reflecting his sustained influence on the fields of robot learning from demonstration, visual object recognition, and embodied humanoid interaction.