HARI: A new approach to quantifying MSD risk factors for hands and fingers

Abstract

Musculoskeletal disorders have become a significant concern for both companies and their employees, with approximately 38% of these MSD impacting the wrists and fingers in France.Despite the essential function of fingers in industrial tasks such as handling objects and assembling parts, there's currently no established ergonomic approach to assess the MSDs risk factors for these body parts.Based on the existing knowledge and theories in physiology, biomechanics, and ergonomics, coupled with technological advancements in computer vision, we propose in this work to introduce a new method to analyze biomechanical risks for the fingers and wrists. In the first part of this work, we develop a markerless approach using MediaPipe Hand Tracking Library to track the motion of the fingers and wrists over time. This library uses machine learning algorithms to detect and track fingers and wrist movements in real time, with low-latency performance.However, it’s important to note that the hand landmarks extracted from MediaPipe are presented in the world coordinates system. So, it’s essential to implement a kinematic model that respects the parent-child relationship and provides reliable joint angles. The kinematic model used in this study includes 25-degree-of-freedom, enabling calculations to be performed in accordance with the hand's functional anatomy. This approach enables us to objectively quantify postural stresses on fingers and wrists during task execution. In the second part of this work, we propose an ergonomic risk assessment method for wrists and fingers based on the type of grip set by the operator that takes into account the different biomechanical factors as well as the aggravating factors. Key factors contributing to ergonomic risk are pinpointed through a review of the existing literature and a survey conducted among diverse companies addressing this type of Musculoskeletal Disorders (MSDs) within their organizational framework. These factors include fingers/wrist posture, exertion frequency, duration per exertion, motion speed, duration per day and force for biomechanical factors. Additionally, aggravating factors like object diameter, object length, visual precision level, vibration, and temperature are all contributors to the development of musculoskeletal disorders. Next, we establish distinct risk levels and set thresholds for each identified risk factor. Evaluators assign ratings to individual variables based on measured or observed exposure data ; then, a multiplier value is assigned to each variable based on existing knowledge and theories across various fields. Using the identified multipliers, the final global score is calculated and interpreted based on the obtained value.A software tool has been developed to automatically process the Hari method with automatic posture assessment. This paper proposes a framework for ergonomic risk assessment by developing a scoring system divided into three levels: red, orange, and green. The proposed scoring system makes it easy for the decision makers to determine whether a given ergonomic risk on fingers and wrists needs to be mitigated or not.A possible extension of this work involves conducting an experimental study to validate the method with occupational safety and health specialists. It is also possible to add functions for automatic calculation of gesture frequency using machine learning algorithms, as well as automatic detection and recognition of the hand-grip gesture.