Subjective and Objective Thermal Comfort Estimation Using Wearable Sensors and Environmental Sensors

An ideal self-adjusting environment requires adapting human thermal comfort automatically and continuously measuring the changes in comfort. According to PMV (Predicted Mean Vote) model, could be evaluated by biometric data environmental data. In this paper, we proposed a method estimate through small number of wearable sensors based on machine learning. There are two typical definitions comfort: subjective representing perception heat objective calculated formula. We used questionnaire formula obtain correct label for kinds comfort, respectively. Among three learning models, random forest has 0.73 MAE, which is suitable estimating 7-level neural network 0.47 investigated estimation accuracy changing sensors' combinations. As result, still roughly