Estimating the relationship among occupant behaviours and indoor environmental parameters using Bayesian networks

This paper investigates the relationship among occupant behaviors and indoor environmental parameters. We collect data from three aspects: indoor environmental variables measured by a wireless sensor network, body sensors carried by occupants, and questionnaires completed by occupiers. The wireless sensor network measures four parameters: temperature, humidity, light and barometric pressure. The body sensors measure collect data including burned calories, physical activity, body position, and galvanic skin response from their wearers. The questionnaire collects data such as occupier’s behaviour and mood. The experiment is conducted in an air conditioned

office where temperature can be controlled by the occupants. Based on such information, we build up a Bayesian network model revealing how indoor environmental parameters and occupant’s behaviour impact physical parameters of human bodies. We also compare our models with Fanger’s thermal comfort model.