“Bend your knees when you pick up a heavy object.” We’ve all been told this so many times that just about everyone should know this by now. Yet workers continue to suffer injuries that are classified as “work-related musculoskeletal disorders” (WMSDs) that result from repetitive movements and inadequate postures. According to one study, more than 40 million workers are affected by WMSDs in Europe alone. Too often, the problematic behaviors aren’t identified until after an injury occurs in spite of training and other programs.
Researchers from the University of Quebec at Chicoutimi have developed a system that can monitor a worker’s movements in real time, and in theory, could issue a warning to the worker to take corrective action. The system is designed to be simple yet accurate. A motion sensor is embedded in a hard hat (which many workers need to wear any way) and a special insole that has four pressure sensors in it. One interesting aspect of the system is that they have found that they only need to put the insole in one shoe, even if it is not the dominant foot. Their research shows that there is a high-correlation between the center of pressure (COP) measurements for either foot, so data from just one is generally sufficient. The motion and force sensors are far less complex and expensive than the force platforms traditionally used to analyze a worker’s movements.
The team used neural networks to analyze the data, and have achieved a 95% accuracy rate in recognizing postures that are likely to result in a WMSD. One bonus is that the system can also detect when a worker is in a given position for a prolonged period of time. Even if it is a “safe” posture, staying in that position for too long can also result in injury. Ultimately, this system could be used to generate alerts to workers (and perhaps supervisors) when they are in a risky posture. The group is planning additional research to evaluate the addition of more sensors, to gather more information such as EMG data from upper arms to detect repetitive motions.
