Microcontroller-based stand-alone alert recgniton device for the hearing impaired using fast fourier transform

• Main Authors: Cai, Carolina L., Gupo, Jason Cydrick V., Urbano, Frances Louise T.
• Format: text
• Language: English
• Published: Animo Repository 2017
• Subjects: Microcontrollers; Electronic controllers; Hearing impaired > Means of communication
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/8362
• Institution: De La Salle University

The effectivity of most devices designed for easier interaction between people and their surroundings is not inclusive when it comes to people with disabilities such as the deaf or hard of hearing (DHH). For the DHH, options are limited when it comes to gadgets involving sound recognition as compared to speech recognition especially those that identify significant warning sounds. Some available notification systems present a disadvantage in that they are mostly wired or transmitter-based systems. This paper presents the development of a portable stand-alone alert recognition device. The designed device can identify high (e.g. fire alarms and sirens) and low priority notification sounds (e.g. doorbells and telephone rings) and can inform the user through visual and vibrational indicators. Collecting most common alert sound and extracting their audio fingerprints trained the device to detect these on the surrounding noise. FFT with spectral peak location was utilized as the core algorithm used in the program. The alert recognition program, audio fingerprints and other hardware components such as microphone, power source, and vibration circuit are integrated to implement the device prototype. Based on the results, the developed device was able to identify the four alert sounds intended in the study with consideration of sound source distance and detection time.