Development of the SENSE - A Smart Wearable Device for Hazard Sound Detection and Vibration Feedback for Deaf Individuals

The SENSE wearable device was developed to enhance safety for individuals with hearing impairments by detecting hazardous sounds and providing real-time vibration feedback. Traditional assistive devices often fail to address real-time sound detection, particularly in noisy environments. This study evaluates the SENSE device, which uses a Custom Convolutional Neural Network (CNN) for sound classification and haptic feedback for alerts. The device was tested on sounds like fire alarms, car horns, and sirens across different noise levels and distances. Results show 92% accuracy in recognizing hazardous sounds with detection times under 4 seconds. However, at greater distances and in high-noise environments, slight inaccuracies in directional prediction were observed. Despite these challenges, the device provided timely haptic feedback within 3.5 seconds. Vibration motor calibration at medium power ensured clear alerts without discomfort. These findings highlight the SENSE device's potential to improve safety for the hearing impaired while suggesting future enhancements in noise filtering, power efficiency, and user customization.