An IoT-driven emergency medical service vehicle detection system using WaveResNet and YOLOv8 for real time traffic management

View at publisher Summarise with VIRAT
Authors: S. G. Shaila, U. Pavan Kumar, S. G. Sumana, Vinu Raj Vamshi, T. M. Rajesh, J. B. S. Dilip Kumar
Cited by: 0
References: 8

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Visual and audible warning signals facilitate operations at high velocities by emergency vehicles. The right of way belongs to other road users. Despite all these precautions, there are situations where drivers can look at emergency vehicles near them and be driving so distracted that traffic incidents and response delays occur. This research presents novel audio and vision based systems to detect emergency vehicles and expedite their passage by modifying traffic signals into green. First, we present the YOLOv8-Emergency Vehicle Detection system, a customized YOLOv8 model especially intended for emergency vehicle detection. At the same time, we build a unique image dataset to improve the vision-based emergency vehicles detection. Our approach to incorporate cross stage partial connections strongly improves the YOLOv8 Emergency Vehicle Detection model. It results in an extraordinary 95% mean average precision exceeding other single stage object detectors. At the same time, we propose WaveResNet, a convolutional neural network designed for the end to end detection of emergency vehicles in audio. The thing that this system does is basically compare the sounds of sirens to the background ‘noise’ of traffic. WaveResNet is robust and accurate above 95% in normal traffic situations using raw waveform inputs of at least one second and noise. Both WaveResNet and YOLOv8-Emergency Vehicle Detection systems also meet real time operating requirements. To bring together a comprehensive solution, we combine the WaveResNet and YOLOv8-Emergency Vehicle Detection systems with an Arduino board to control traffic signals. Solving the emergency vehicle detection problem, this innovative method combines vision and audio based technologies to optimize real time response and traffic signal control.

Fields of Study & Subjects

Advanced Neural Network Applications IoT and GPS-based Vehicle Safety Systems Autonomous Vehicle Technology and Safety

Publication details

Journal / source
AIP conference proceedings
Volume
3401
Pages
020035–020035
Publication date
1 January 2026
Language
EN