An Ultrareliable Low Computational Method for Detecting the Direction of Emergency Vehicle Sirens Based on an Array of Microphones

Autonomous driving is becoming more and more advanced and soon it can represent a real solution for reducing traffic accidents and as a result traffic deaths. In the present paper, we present an emergency vehicle direction of arrival detection backup system for the standard detection systems based on computer vision and/or V2X communication technologies. The proposed system is ultrareliable working for a SNR of under -50 dB and it has a low computational complexity compared to the general available methods. Also, the present system can be used to help drivers identify emergency vehicle sound direction of arrival, especially in cars with performant soundproof interiors.