Emergency Siren Detection System for Deaf People

Because of increasing traffic density, emergency vehicles may pose a threat to people with hearing difficulties. This work aims to develop a lightweight system which can be worn as a bracelet, to detect the acoustic emergency signals from these vehicles and alert the user of their presence. After the signal is captured by a microphone, it is processed using analog blocks and a repetitive series of pulses are obtained. The analog processing stage was simulated using electronically generated siren tones and self-made street recordings in Cluj-Napoca city as input. The pulses are then processed by microcontroller which decides if the signal is present. The digital processing stage was implemented on an ATmega328P microcontroller and simulated with the signals obtained from the previous stage as input. Simulations showed that the system can detect most recordings. A PCB layout was also designed and implemented. Tests on the PCB proved that the analog processing blocks are working properly.

I. Introduction

Recent developments to detect emergency vehicles (police, ambulance, fire-fighting trucks) that are based on the acoustic signal emitted by the vehicle, also known as a siren, exist in the field of audio signal processing.