Preliminary Development of a Gesture Recognition System for Posture Monitoring in Chainsaw Cutting: A Step Towards Enhanced Safety for Forestry Workers

Hand-Arm Vibration (HAV) is a prevalent occupational hazard affecting millions of workers, leading to various health issues. Among them, Musculoskeletal Disorders (MSDs) are frequently observed in workers who engage in repetitive tasks and adopt awkward postures. Training workers and monitoring their fatigue levels are essential measures to prevent injuries and the onset of occupational diseases. This study contribute to the first stage in the developing a gesture recognition system using Inertial Measurement Units (IMUs) for posture and movements monitoring in chainsaw cutting. Experiments were conducted simulating different cutting operations, and acquiring signals of body joint rotations. An algorithm combining autoencoder and random forest classifier was developed to classify the cutting conditions leading to an accuracy close to 95%, even when using just four sensors. Despite the limited data available, the findings are promising for future research into how injuries or occupational diseases affect the way forestry workers use chainsaws or other hand-held powered tools.