Smart monitoring of worker posture in an office environment

Doing computer work with an incorrect posture can lead to musculoskeletal disorders. Training workers with ergonomics advice decreases these risks. In this paper a posture recognition system, CAPRIO, is proposed, using pressure sensors and context information to generate tailored ergonomics advice. Using pressure sensors alone leads to too many false positives in detecting wrong postures. Using computer usage as an additional input improves the accuracy of the posture detection. The system is compared by simulation to other possible methods on accuracy, implementation complexity, usability and privacy sensitivity. CAPRIO scores well on those criteria. It has lower accuracy than some video based methods, but is less privacy infringing.