 |
Prof. Binrui Wang Vice President China Jiliang University |
Biography
Prof. Binrui Wang is a doctor, professor, doctoral supervisor and the vice-president of China Jiliang University. He received the Ph.D. degree in pattern recognition and intelligent system from the School of Information Science and Engineering at Northeastern University in 2005. His main research areas include bionic robotics, intelligent perception and metrology. He has published over 200 high-level academic papers, authored 3 monographies and 2 textbooks, held over 50 authorized invention patents as the first inventor, and developed 7 national standards. He has received the second prize from the Chinese Society of Automation and the second prize from the China Instrument and Control Society. He serves as the vice chairman of the Robotics Professional Committee of the China Association for Standardization, a member of the National Standards Committee Technical Committees TC591 and TC307, the vice secretary-general of the National Civil Aviation Metrology Technical Committee, and member of the Space Metrology Technical Committee.
Title
Neuromusculoskeletal Motion Control for Humanoid Robots
Abstract
Pneumatic artificial muscles (PAMs), characterized by a high power-to-weight ratio and inherent compliance, provide a promising actuation approach for compliant and efficient humanoid robot motion. However, their intrinsic nonlinearities, together with the high redundancy and strong coupling of multi-muscle humanoid systems, pose challenges to stable, coordinated, and adaptive motion control. Focusing on anthropomorphic motion control for PAM-actuated humanoid robots, this keynote presentation introduces a neural–muscular–skeletal inspired control framework. The framework integrates rhythmic control based on the central pattern generator (CPG) mechanism, learning-based parameter adaptation, and bio-inspired reflex regulation to coordinate multiple pairs of antagonistic PAMs. Specifically, CPG neurons generate rhythmic locomotor patterns, while deep reinforcement learning (DRL) serves as an upper-level brain-like learning mechanism to optimize key CPG parameters. Stretch and vestibular reflex mechanisms are further incorporated as fast sensory feedback pathways for impact rejection, muscle tension regulation, and posture adjustment. Through the integration of rhythm generation, learning adaptation, and reflex feedback, the framework aims to enhance gait coordination, motion stability, and environmental adaptability in complex motion states and under external disturbances.