Sign in Subscribe
Robotics industry

Breakthrough in Prosthetic Control: Researchers Develop Neural-Driven Approach for Simultaneous Gestures and Forces Recognition

Strategic Insights

— 1 min read

Researchers at Zhejiang University have made a groundbreaking discovery in the field of robotics and automation, presenting a novel approach for simultaneous recognition of gestures and forces in prosthetic control. The study, published in Ieee Robotics and Automation Letters, introduces a neural-driven simultaneous recognition approach based on the motor unit (MU) discharges, which enables flexible interaction between upper limb amputees and their environment. This innovative method has demonstrated exceptional accuracy, with average recognition rates of 96.26% for gestures and 99.92% for forces.

Key Takeaways:

  • The researchers proposed a neural-driven simultaneous recognition approach based on the motor unit (MU) discharges, using high-density surface electromyogram (HD-sEMG) signals and the geodesic flow kernel (GFK) algorithm to enhance neural-driven features.
  • The approach uses two random forest (RF) classifiers to predict gestures and force levels from enhanced neural-driven features, demonstrating superior accuracy compared to conventional sEMG methods.
  • The study tested the proposed method on eleven gestures and three force levels, achieving average recognition accuracies of 96.26% for gestures and 99.92% for forces.
  • The geodesic flow kernel (GFK) algorithm improved the classification accuracy of gestures and forces by 0.98% and 3.5%, respectively.
  • The researchers recognized the potential of this approach to enable natural and intuitive human-machine-environment interaction, promoting the intelligent upper limb prosthesis to precisely and simultaneously control gestures and forces.
  • The study was financially supported by the National Natural Science Foundation of China (NSFC), National Key Laboratory of Transient Impact, Key Research and Development Programme of Zhejiang, Scientific Research Fund of Zhejiang University, and Major Scientific and Technological Achievements Transformation Project of Hebei Province.

Statistics:

  • Average gesture recognition accuracy: 96.26%
  • Average force recognition accuracy: 99.92%
  • Improvement in classification accuracy of gestures: 0.98%
  • Improvement in classification accuracy of forces: 3.5%

Sources:

  • Neural-driven Simultaneous Recognition of Gestures and Forces Via Feature Enhancement for Prosthetic Control. Ieee Robotics and Automation Letters, 2025;10(11):11331-11338.
  • Tao Liu, Zhejiang University, School of Mechanical Engineering, State Key Lab Fluid Power & Mechatron Syst, Hangzhou 310027, People's Republic of China.

Read more