Intelligent Arduino Streetlight System for Environmental Monitoring and Accident Detection

Abstract

Smart urban infrastructure plays a significant role in improving the safety of the roads while also lowering the energy usage. In an ideal world, efficient accident detection systems in place, there is a timely response from an emergency service which would prevent loss of lives and properties during accidents, both in rural and urban areas. The aim of this project is to provide a solution to these problems by developing a system that is effective from a cost point of view and provides combination of environment monitoring in real time and impact of the accident. This paper presents an intelligent streetlight management system developed using Arduino Mega as a main controller along with environmental sensors including SHT31 temperature and humidity sensor, GL55 photoresistor (LDR), and ultrasonic sensor to help it monitor environmental conditions and only work when needed keeping it energy efficient. Real-time detection led to transfer of a model based on YOLOv8, which was built and trained using 1,807 photographs of individuals either being or not being hit; all this data was merged towards an ESP32 camera for on-the-go video analyses. They can then send accident data -- times, dates, GPS locations, whatever -- to a dashboard for instant response. Notably, the streetlight system operates solely under predetermined conditions, conserving energy, and the accident detection system reached a mAP@50 score of 0.6305 and a mAP@50–95 score of 0.4342, achieving a 60% accuracy when applied to video footage taken in real-world scenarios. These outcomes clearly show the viability of the integration of environmental observation and protection systems within a single low-cost unit. This project demonstrates how Arduino-based systems can help solve urban problems, generate energy savings, increase road safety, and identify sustainable solutions for urban growth.