Max Robotics — Global Service Robot Directory + US Certification

About

Affan Pervez’s research lies at the intersection of robot skill learning, programming by demonstration, and human-robot interaction, with a particular focus on making robots more adaptable and intuitive to teach. His most influential work centers on advancing **Dynamic Movement Primitives (DMPs)** — a core method for encoding motion data. Pervez pioneered the use of **mixture models** and **convolutional neural networks** to learn task-parameterized DMPs, enabling robots to generalize learned skills to new situations without relying on custom vision systems. This breakthrough, detailed in his highly cited 2017 papers (76 and 53 citations respectively), directly addressed a key bottleneck in real-world deployment. He also made significant contributions to **learning from demonstration for teleoperation**, developing methods to reduce operator mental workload by learning repetitive tasks from human demonstrations — a contribution cited over 50 times. His work on **robotic calligraphy**, learning motor skills directly from character images, showcases his ability to bridge perception and action. With over 200 total citations, Pervez’s research has helped shape modern approaches to task-parameterized skill acquisition, making robot programming more accessible and robust for real-world applications.