InstantRR: Instantaneous Respiratory Rate Estimation on Context-Aware Mobile Devices

Chronic obstructive pulmonary disease (COPD) is one of the leading causes of death both in the USA and worldwide. Respiratory rate is an important predictor for acute COPD exacerbation and an indicator of overall well-being for healthy individuals. Current methods to measure respiratory rate either involve uncomfortable, specialized sensors such as a chestband or are less resilient to varying real-life situations. In this paper, we present a novel context-aware framework that can reliably estimate respiratory rate using data from sensors embedded in users’ existing mobile devices such as smartphones and smartwatches. Our approach takes current contexts, such as device placement, user’s social interaction, and user’s pulmonary health condition into consideration, and finds the optimal fusion across sensor streams and algorithms. We show that our approach can handle varying user contexts (e.g., detecting device placement with an accuracy of 97%) and reliably estimate respiratory rate with errors as low as 0.85 breaths per minute.