Vladimir Joukov

University of Waterloo, Harbin Engineering University, Sepuluh Nopember Institute of Technology

Papers

Total Citations

H-Index

About

Vladimir Joukov’s research lies at the intersection of human motion analysis, imitation learning, and human-robot interaction, with a focus on enabling robots to understand and replicate human behavior. His work addresses fundamental challenges in how machines perceive and learn from continuous, real-world human movement data. A key contribution is his development of a full-body multi-primitive segmentation method using classifiers, which allows robots to automatically break down continuous motion streams into meaningful action units—a critical step for applications like gesture recognition and imitation learning. This work has garnered 11 citations and established a foundation for more sophisticated interaction systems. Joukov further advanced the field by introducing a Gaussian process-based model predictive controller for imitation learning, enabling robots to infer and minimize task-specific cost functions from human demonstrations without restrictive assumptions. This approach, cited 6 times, bridges the gap between human motor control principles and robotic learning. Additionally, he has contributed to wearable sensor technology, proposing a Generalized Hebbian algorithm for estimating sensor-to-body orientation in inertial measurement units, a method with direct applications in exoskeletons and active prosthetics. Through these contributions, Joukov has helped shape how robots perceive, segment, and learn from human motion, advancing the frontier of intuitive human-robot collaboration.