Design of Low-Power System for Detection of Abnormal Human Audio
- February 23, 2018
- Posted by: RSIS
- Categories: Electrical and Electronics Engineering, Electronics & Communication Engineering, Engineering
International Journal of Research and Innovation in Applied Science (IJRIAS) | Volume II, Issue X, December 2017 | ISSN 2454-6194
Design of Low-Power System for Detection of Abnormal Human Audio
G. Bhavana1, C.Bhargav2, T.Chakrapani3, K .Sudhakar4
1M.Tech, VLSI, Dept of ECE, ST. Johns College of Engineering and Technology, Kurnool, Andhra Pradesh, India.
2Assistant Professor, Dept of ECE, ST. Johns College of Engineering and Technology, Kurnool, Andhra Pradesh, India.
3Associate Professor, Dept of ECE, ST. Johns College of Engineering and Technology, Kurnool, Andhra Pradesh, India.
4H.O.D, Associate Professor, Dept of ECE, ST. Johns College of Engineering and Technology, Kurnool, Andhra Pradesh, India
Abstract: We present a low-power, efficacious, and scalable device for the detection of symptomatic patterns in human audio signals. The digital audio recordings of quite a number symptoms, such as cough, sneeze, and so on, are spectrally analyzed the usage of a discrete wavelet transform. Subsequently, we use simple mathematical metrics, such as energy, quasi-average, and coastline parameter for a number wavelet coefficients of interest relying on the kind of pattern to be detected. Furthermore, a mel-frequency cepstrum-based analysis is applied to distinguish between signals, such as cough and sneeze, which have a similar frequency response and, hence, manifest in common wavelet coefficients. Algorithm-circuit codesign methodology is utilized in order to optimize the gadget at algorithm and circuit ranges of graph abstraction. This helps in imposing a low-power system as properly as preserving the efficacy of detection. The gadget is scalable in phrases of consumer specificity as well as the kind of sign to be analyzed for an audio symptomatic pattern. The proposed structure of this paper analysis the logic size, location and power consumption using Xilinx 14.2
Keywords: digital audio recordings, cepstrum-based analysis, Voltage scaling, Area, efficacy.
I. INTRODUCTION
These systems monitor various internal in addition as external parameters associated with the human health, like temperature, heart rate, and so on. Apart from these parameters, it’s accepted that symptoms, like cough, sneeze, belching, and so on. are early markers of great health problems, like respiratory illness, diarrhea, and respiratory illness, particularly among kids. If repetitive prevalence of those symptoms is detected in advance, it’s potential for the patient or the health care personnel to begin remedial action before aggravation of the issue.
