Design and Fabrication of a Laboratory Model of Physical Workload Classification Tool and Evaluation of Its Usability in the Real Work Environment

%VO2max index is the gold standard of physical workload classification. Because of difficulty of this index measurement, it is not a practical method in the real work places. Recently, a new method was proposed for estimating %VO2max through three parameters of resting heart rate, working heart rate, and weight based on a neuro-fuzzy network in MATLAB software. The goal of this study was designing and fabricating a laboratory model of physical workload classification tool based on the mentioned method. The programming of the device was performed with the Arduino software and in C++ language in the AVR microcontroller; then, it was entered into integrated circuit (IC) by the programmer. The output of heart rate sensor was entered into the microcontroller through I2c protocol. The usability score of the device was evaluated by 20 occupational health experts employed in the industry and was compared with manual physical workload classification. Findings: The mean usability score of this system was 84.6 ± 7.3 and was ranked in B category. It means that the usability of the system is very good. The required time for physical workload classification using this tool was approximately half of the required time for this work without the tool. The adaptive neuro-fuzzy inference system (ANFIS) was presented for estimating physical workload in the form of a practical tool in order to be used in the industry. Regarding the lack of sufficient accuracy of current indexes for estimating physical workload such as heart rate, this fabricated tool is a proper substitute for the former methods. High usability and low required time are two main advantages of the proposed tool