People with speech disorders, such as Parkinson’s disease, oftentimes struggle to take part in normal conversation due to a cerebral disconnect between their actual and perceived speech noise level. Offering a device to provide continuous speech therapy would assist the patient in everyday conversation, ensuring a proper speaking level is met through real-time data processing.
The team will design and prototype a head-mounted listening device that monitors the amplitude and spectral density of the wearer’s voice and compares it to a designated criterion. The device will provide real-time haptic feedback when an amplitude threshold is not met, as well as record amplitude data to plot trends of the user’s speech activity. The headset will consist of two micro-electrical mechanical systems (MEMS) microphones and one accelerometer to monitor the user’s voice even in high noise scenarios. All audio processing will be conducted through an embedded digital signal processor (DSP) in a wired belt pack. The DSP will analyze the received signals from the microphones, reduce ambient noise, and compare the user’s voice to their designated threshold.