Human Posture Estimation: A Systematic Review on Force-Based Methods-Analyzing the Differences in Required Expertise...

Area of Science:

• Biomechanics
• Human-Computer Interaction
• Sensor Technology

Background:

• Camera-based human motion capture is often impractical.
• Force-based human posture estimation (FPE) presents an alternative for sensor integration in smart products.
• Existing FPE methods lack a consolidated overview, hindering interdisciplinary research.

Approach:

• A systematic review following the PRISMA 2020 guidelines was conducted.
• 82 studies were selected, categorizing methods into machine learning (ML)-based (59) and digital human model (DHM)-based (23).
• Analysis focused on required expertise, benefits, and accuracy of each method.

Key Points:

• ML-based FPE utilizes sensor data (primarily pressure mapping) and trained models, achieving over 90% accuracy.
• DHM-based FPE models human body structure and kinematics to minimize physical stress, requiring human physiology expertise.
• ML methods need application-specific training but less domain expertise, while DHMs offer general applicability.

Conclusions:

• Both ML-based and DHM-based FPE methods have distinct advantages and disadvantages.
• The choice of method depends on application-specific needs and available expertise.
• This review provides a foundation for utilizing FPE in patient monitoring, ergonomics, and sports science.