Toward Real-Time Detection of Object Lifting Using Wearable Inertial Measurement Units

Area of Science:

• Occupational Health and Safety
• Biomechanics
• Human-Computer Interaction

Background:

• Workers performing strenuous physical labor face high risks of back and other occupational injuries.
• Assistive devices that detect lifting motions in real-time could mitigate these risks.
• Accurate and immediate detection of lifting onset is crucial for effective assistive device operation.

Purpose of the Study:

• To develop and validate a real-time system for detecting the start and end of object lifting motions.
• To enable the development of assistive devices that can provide immediate support during physical labor.
• To reduce the incidence of occupational injuries associated with manual lifting tasks.

Main Methods:

• A novel algorithm was developed to detect lifting motions in real-time.
• The system utilizes pelvic angle and the distance between the user's hands and their center of mass as input.
• An algorithm was designed to identify hand-center distance peaks and verify lift occurrence using pelvic displacement.
• The system was tested on 5 participants performing 100 lifts of varying types.

Main Results:

• The developed system demonstrated reliable real-time detection of lifting motions.
• The median time error for detected lifts was 0.11 seconds.
• 95% of detected lifts occurred within 0.28 seconds of the actual lift.
• The system achieved high accuracy in identifying both the start and end of lifting actions.

Conclusions:

• Real-time detection of lifting motions is feasible using pelvic angle and hand-to-center of mass distance.
• This technology can support the development of advanced assistive devices for physically demanding jobs.
• The findings suggest a significant potential for reducing occupational injuries through smart assistive technologies.