ESE498: Electrical Engineering Capstone Project
Ari Cotler, Lydia Reader, Mo Wu
Under advisement of Dr. Jason Trobaugh, Dr. Jim Feher, and Dr. A.U. Bankitis (Oaktree Products)
Non-speech sound recognition is an important branch of hearing assistive technologies. The goal of our research is to make an affordable device for the hearing-impaired community that alerts the user of urgent household tones such as doorbells, smoke alarms, and appliances.
We designed and implemented an algorithm for a collection of common household tones (two doorbells, smoke alarm, and microwave chime) on a microprocessor to show the possibility of developing a portable device for sound recognition. The algorithm is implemented on an Arduino microprocessor with a vibration motor and an LCD monitor to inform the hearing impaired with haptic feedback and a text display. Our prototype is able to detect the four tones with high accuracy and minimal delay under controlled noise.
We chose the Arduino platform for the prototype. There are three sub-circuits: the microphone, vibration motor, and LCD screen.