Abstract
Current evolutions in the Internet of Things and cloud computing make it believable to build smart cities and homes. Smart cities provide smart technologies to residents for the improved and healthier life, where smart healthcare systems cannot be ignored due to rapidly growing elderly people around the world. Smart healthcare systems can be cost-effective and helpful in the optimal use of healthcare resources. The voice is a primary source of communication and any complication in the production of voice affects the personal as well as professional life of a person. Early screening of voice through an automatic voice disorder detection system may save life of a person. In this paper, an automatic voice disorder detection system to monitor the resident of all age group and professional backgrounds is implemented. The proposed system detects the voice disorder by determining the source signal from the speech through the linear prediction analysis. The analysis calculates the features from normal and disordered subjects. Based on these features, the spectrum is computed, which provided distribution of energy in normal and voice disordered subjects to differentiate between them. It is found that lower frequencies from 1 to 1562 Hz contributes significantly in the detection of voice disorders. The system is developed by using sustained vowel and running speech so that it can be deployed in a real world. The obtained accuracy for the detection of voice disorder with the sustained vowel is 99.94% ± 0.1, and that is for running speech is 99.75% ± 0.8.