Technologies for developing ambulatory cough monitoring devices

Cough is a prevailing symptom in most lung diseases. While cough sounds themselves can be very instrumental in the diagnosis of certain diseases, their intensity and frequency also infers the intensity of the particular illness. There is an imperative need for a robust system for identifying and analyzing cough sounds. In implementing such systems, researchers are confronted with technical challenges such as the choice of sensors and methods of signal acquisition, the real time analysis of the acquired signals, and the accurate identification of cough events, distinguishing them from similar sounds such as speech, laughing, throat clearing and sneezing. Previous approaches have employed external environmental sensing methods to achieve more accurate detections at the expense of mobility, scalability and real-time cough sensing. Alternative approaches have proposed wearable cough sensing methods, which, while mobile, can often face challenges in terms of robustness and obtrusiveness. In this paper, we explore the strengths and shortcomings of the various techniques that have been proposed for automatic detection and analysis of cough sounds. We also suggest the next steps in furthering the state of the art.