Desarrollo e implementación de un prototipo de robot explorador de inspección horizontal con control manual y visión artificial para detección de fallas en infraestructuras cableadas

Abstract

This research focuses on the use of an exploratory robot for the horizontal inspection of electrical and telecommunications wiring located in confined spaces. The proposed prototype is aimed at hard-to-access areas, such as false ceilings, where manual inspection is usually slow and limited by environmental conditions. The system uses a mobile robotic platform manually controlled through LoRa wireless communication and an architecture based on ESP32 microcontrollers. For visual inspection, a HuskyLens sensor trained with different types of wiring faults was used. During the robot’s operation, images and video were captured using an ESP32-CAM module. The collected information was stored on a local server implemented in Node.js. The tests were carried out inside a residential building and in one of the laboratories of the Universidad Politécnica Salesiana, specifically in a false ceiling. The environment presented typical conditions of this type of space, such as low lighting, dust presence, and structural elements that hinder movement. At this stage, six types of wiring faults were considered. For system evaluation, tests were conducted at three distances (15 cm, 30 cm, and 45 cm), using a total of 360 images obtained during the experiments. The results showed that system accuracy decreases as the inspection distance increases, reaching an average of 85% at 15 cm, 75% at 30 cm, and 65% at 45 cm. In addition, the system maintained stable data transmission, allowing proper reception and storage of information. During this process, a higher number of faults were detected, demonstrating the effectiveness of the system under real operating conditions.