Edge AI and IoT-Based Smart Helmet for Real-Time Hazard Detection and Safety Monitoring of Drainage and Sewer-Line Workers

Workers who clean and inspect underground drainage and sewer lines face toxic and flammable gases, heat, humidity and poor ventilation, usually far from anyone who could help. Ordinary safety helmets protect only against physical impact and cannot sense or report these dangers. This paper presents the design of a wearable smart helmet that combines three sensors—an MQ-136 hydrogen sulfide sensor, an MQ-4 methane sensor and a DHT11 temperature-humidity sensor—with an ESP32 controller, an Artificial Neural Network (ANN) running on the device, and a cloud-linked supervisor dashboard. The ANN takes the four sensor readings as input and classifies the working environment as Safe, Warning or Danger. The result drives a buzzer and red and green LEDs for the worker, while the ESP32 uploads status to the cloud over Wi-Fi. A fixed threshold rule runs beside the ANN so that a clearly unsafe reading can never be down-graded by the model. The paper describes the architecture, the role of the ANN, the alert logic and a planned evaluation of the prototype