인공지능 스마트헬멧을 이용한 근접경보시스템 개발

Abstract

Smart helmet refers to a helmet with a smart function that integrates sensors, communication, and actuator modules, not a general helmet. In this study, the smart helmet combined with ESP 32-CAM, a Camera module with an Arduino Uno board and Wi-Fi communication function, to exchange images with a smartphone through an app inventory, and analyzed HTTP communication with a public artificial intelligence open platform, Etri server. We developed a Proximity Warning System (PWS) that detects objects in the image through object detectors and receives data again on smart phones, signals Arduino through Bluetooth communication on smart phones, and alerts workers and drivers through LED straps and tricolored LEDs. Performance experiments were conducted on systems developed in outdoor environments by distance, and subjective workloads were evaluated through the Nasa-TLX survey to evaluate the subjective workload felt when workers used them. In this study, a proximity warning system using an artificial intelligence smart helmet was developed and implemented and utilized in three environments: a two-wheeled vehicle driver, a construction site, and a mining site. The proposed personal PWS has the advantage of being able to directly provide visual proximity alerts to both workers and drivers, allowing them to quickly identify and evacuate dangerous situations and provide visual alerts to pre-registered vehicles or equipment.