Towards non-intrusive metabolic rate evaluation for HVAC control

摘要: To provide comfortable environments by heating, ventilation, and air conditioning (HVAC) systems, occupants’ perspective should be taken into account. The predicted mean vote (PMV) model, which is commonly used in design of HVAC systems, accounts for human factors such as metabolic rate. However, pre-determined values (ie, met unit) for metabolic rates are used due to difficulty in their contextual and real-time evaluation. As a solution, we propose a novel and non-intrusive respiration ratio evaluation method using a cost-effective Doppler radar sensor (DRS) system. This method has been inspired by the statement from ASHRAE: respiratory ratio (ie, the ratio of exhalation to inhalation) is correlated with heat production of the body (ie, metabolic rate). DRS systems could be used to identify subtle movements, including periodic movements such as pulmonary activities, by processing the modulated returned signal from an environment. Therefore, using the information, contained in the return signal, exhalation (𝑄𝐶𝑂2 exhaled) and inhalation (𝑄𝑂2 inhaled) can be distinguished by processing transitions from local minima to maxima to minima, which indicate inhalation and exhalation, respectively. We leveraged the representation of pulmonary activities (intensity and duration) in the DRS signal to calculate respiration ratio. Our proposed framework uses the Savitzky-Golay method and band-pass filtering to reduce noise in the time-domain representation of the signal and utilize the peak detection algorithm to distinguish each pulmonary activity. By multiplying intensity and duration of each pulmonary activity the volumetric flow of inhalation …