- This article examines how the application of AI-based 3D video capture technology can be used to help with predicting shoulder injury risk.
- It provides a brief explanation of the difference between 2D and 3D video capture technology and explains why AI-based technology needs accuracy and validation, especially if used for injury prediction and prevention.
- It also explains the process used to adapt the ACGIH upper limb localized fatigue threshold limit value for practical utilization in the shoulder injury risk assessment.
Most ergonomic risk assessments are completed by watching a person work, making measurements, performing calculations, and comparing the results against a checklist, index or published standard. In addition to being time-consuming, these observational assessments are prone to inaccuracy and inconsistency due to interoperator variance. If 10 different practitioners perform an ergonomic risk assessment on the same job using observational techniques, it is unlikely that they would produce identical results. The results would vary because everyone perceives what they see a little differently when using the standard observational methods deployed today. This is an ideal opportunity to apply artificial intelligence (AI) to objectively measure the performance of a work task and assist the practitioner with ergonomic risk assessments.
This article focuses on the use of AI for identification of ergonomic risk associated with upper-limb fatigue in jobs requiring raised and extended arm postures. AI is increasingly applied to real-world problems in industry to perform tasks that require visual perception, speech recognition, decision-making and in-process quality checks. Advances in AI are combined with advances in mobile devices such as smartphones to incorporate processing power, advanced cameras and inertial measurement unit sensors. Inertial measurement unit sensors are worn on the body to measure and report acceleration and angular velocity as a person’s limbs are in motion.
This article explains the use of AI to characterize a worker’s 3D motion to assist practitioners in providing accurate, repeatable and objective assessments. Without AI assistance, comparable upper-limb fatigue studies would take a practitioner countless hours to complete and require potentially error-prone subjective decisions. This article also describes the enabling technologies, background research and development of a new shoulder injury risk assessment (SIRA) methodology for ergonomic analysis.
