The devices that are been produced by the center sense physiological as well as environmental factors at varying rates. In order to enable any inference tasks (such as detection of asthma exacerbation or a cardiac condition), this data needs to be aggregated in order to enable the extraction of any necessary biomarkers (e.g., heart rate variability, blood pressure, Ozone in the environment). Continuous transmission of all the modalities to an aggregator (e.g., a smartphone) can impact the battery life of the device.
In order to address these issues, data-driven approaches are been developed to identify the context under which specific sensing modalities are needed for inference. This will allow us to develop techniques that can be used for power minimization of the system by determining which signals are needed and at which frequency based on the current state of the user, its environment and the devices.