Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A perspective on neuromechanical biomarkers for neurorehabilitation: towards reliable assessment in research and clinical practice.

Progress in biomedical engineering (Bristol, England)·2026
Same author

A bioinspired hybrid robotic joint for safer physical human-robot interaction.

Bioinspiration & biomimetics·2026
Same author

Data-driven mouse motor thalamus model reveals topography and spatial weight scaling govern spindle dynamics.

Communications biology·2026
Same author

APOLLO11: a bio-data-driven model for clinical and translational research in lung cancer.

NPJ precision oncology·2026
Same author

Are social robots more interesting than humans? Quantitative assessment of Joint Attention in autistic and typically developing children.

Asian journal of psychiatry·2026
Same author

Effect of EMG-activated vibrotactile biofeedback on skill learning in children with genetic and acquired dystonia during point-to-point and cyclic task practice.

Journal of neuroengineering and rehabilitation·2025

Related Experiment Video

Updated: Aug 13, 2025

The Muscle Cuff Regenerative Peripheral Nerve Interface for the Amplification of Intact Peripheral Nerve Signals
07:30

The Muscle Cuff Regenerative Peripheral Nerve Interface for the Amplification of Intact Peripheral Nerve Signals

Published on: January 13, 2022

2.1K

IMU-based human activity recognition and payload classification for low-back exoskeletons.

Mattia Pesenti1, Giovanni Invernizzi2, Julie Mazzella2

  • 1Department of Electronics, Information and Bioengineering, Nearlab, Politecnico di Milano, 20133, Milan, Italy. mattia.pesenti@polimi.it.

Scientific Reports
|January 21, 2023
PubMed
Summary

This study introduces a deep learning system using inertial sensors for robotic exoskeletons. It enables human activity recognition and adaptive payload compensation, enhancing worker safety and exoskeleton performance.

More Related Videos

Training Persons with Spinal Cord Injury to Ambulate Using a Powered Exoskeleton
09:46

Training Persons with Spinal Cord Injury to Ambulate Using a Powered Exoskeleton

Published on: June 16, 2016

20.7K
Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis
11:16

Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis

Published on: July 22, 2014

16.3K

Related Experiment Videos

Last Updated: Aug 13, 2025

The Muscle Cuff Regenerative Peripheral Nerve Interface for the Amplification of Intact Peripheral Nerve Signals
07:30

The Muscle Cuff Regenerative Peripheral Nerve Interface for the Amplification of Intact Peripheral Nerve Signals

Published on: January 13, 2022

2.1K
Training Persons with Spinal Cord Injury to Ambulate Using a Powered Exoskeleton
09:46

Training Persons with Spinal Cord Injury to Ambulate Using a Powered Exoskeleton

Published on: June 16, 2016

20.7K
Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis
11:16

Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis

Published on: July 22, 2014

16.3K

Area of Science:

  • Robotics
  • Human-Computer Interaction
  • Machine Learning

Background:

  • Work-related musculoskeletal disorders, particularly low-back pain, significantly impact industrial workers, leading to absenteeism.
  • Robotic exoskeletons offer potential solutions for improving worker health and quality of life, but current systems face limitations in adaptability.
  • Sub-optimal control systems hinder the widespread adoption of exoskeletons in dynamic occupational settings due to a lack of wearer and task adaptation.

Purpose of the Study:

  • To develop a deep learning approach for industrial exoskeletons to recognize human activities and adapt to varying payloads.
  • To enhance the symbiotic human-robot interaction by enabling exoskeletons to respond dynamically to user actions and environmental demands.
  • To improve the practical application of powered exoskeletons in occupational scenarios through advanced control algorithms.

Main Methods:

  • Utilized inertial measurement units (IMUs) for data acquisition, which are easily integrated into industrial exoskeletons.
  • Employed Long-Short Term Memory (LSTM) networks for both human activity recognition and classification of lifted object weights (up to 15 kg).
  • Conducted subject-specific model training and testing with 12 healthy volunteers and performed a real-time in-lab test in a simulated industrial scenario.

Main Results:

  • Achieved a median F1 score of [Formula: see text] for human activity recognition.
  • Obtained a median F1 score of [Formula: see text] for payload estimation (object weight classification).
  • Demonstrated the successful real-time applicability of the developed deep learning approach in a simulated occupational setting.

Conclusions:

  • The proposed deep learning framework effectively enables industrial exoskeletons to perform human activity recognition and adaptive payload compensation using inertial sensors.
  • The developed algorithms show significant promise for enhancing the adaptability and effectiveness of powered exoskeletons in real-world industrial applications.
  • This approach contributes to advancing symbiotic human-robot interaction, paving the way for more intelligent and responsive exoskeleton systems.