A human-centered framework for data-driven anthropometric sizing: design and validation in a military context

Area of Science:

• Human-Computer Interaction
• Ergonomics
• Applied Machine Learning

Background:

• Traditional anthropometric methods for equipment personalization are expensive, slow, and not scalable.
• High-stakes professions require accurate and efficient personalized equipment fitting.
• Existing methods struggle to accommodate diverse body types and resource constraints.

Purpose of the Study:

• To propose and validate a low-cost, human-centered framework for personalized equipment fitting.
• To integrate machine learning with usability evaluation for a data-driven sizing model.
• To develop and test a smartphone-based prototype for real-world applicability.

Main Methods:

• Applied clustering algorithms to anthropometric data for a data-driven sizing model.
• Developed a smartphone-based prototype for framework validation.
• Conducted usability evaluations using System Usability Scale (SUS) and Questionnaire for User Interaction Satisfaction (QUIS) with university and Air Force Academy students.

Main Results:

• Achieved high usability scores: average SUS of 83 (82.5) and QUIS of 209.55 (187.33).
• Validated the data model, demonstrating effective stratification of complex body types (p < .001).
• Confirmed the framework's success in real-world applicability and user acceptance.

Conclusions:

• The proposed framework offers a generalizable solution for personalized fitting systems in resource-constrained environments.
• Machine learning integration with usability evaluation enhances the efficiency and scalability of anthropometric personalization.
• The human-centered approach ensures user acceptance and practical implementation of personalized equipment solutions.