CBRRES: a computer-based REBA and RULA evaluation system for musculoskeletal disorder estimation.

Area of Science:

• Digital ergonomics
• Human-computer interaction
• Occupational health and safety

Background:

• Traditional ergonomic risk assessments are often time-consuming and subjective.
• Accurate posture assessment is crucial for preventing musculoskeletal disorders.
• There is a need for automated, real-time solutions in ergonomic evaluations.

Purpose of the Study:

• To design and develop a computerized posture assessment system.
• To calculate Rapid Entire Body Assessment (REBA) and Rapid Upper Limb Assessment (RULA) scores.
• To improve the efficiency, accuracy, and provide real-time feedback for ergonomic risk evaluations.

Main Methods:

• Utilized Python and MediaPipe for pose estimation and landmark extraction.
• Developed distinct approaches for upper body and entire body posture estimation.
• Leveraged pre-trained machine learning models for skeletal landmark identification.

Main Results:

• Achieved posture analysis and REBA/RULA scoring within 500-700ms per frame for real-time feedback.
• Attained an average accuracy of 93% in estimating joint angles and ergonomic scores compared to human experts.
• Demonstrated robustness across various postures and lighting conditions.

Conclusions:

• The developed system provides a reliable, automated solution for ergonomic risk assessment.
• This work represents a significant advancement in the field of digital ergonomics.
• The system's accuracy, low latency, and ease of integration offer potential for real-time ergonomic monitoring and intervention.