Autonomous Aquatic Laser-Following Robot Through RGB Sensors and Optimized Artificial Neural Networks

Abstract

Aquatic Robots have a critical role to enhance oceanography studies, enable search and rescue scenarios, and basically enable performing tasks that without them, would be too dangerous or even impossible for humans alone. Among the different types of Aquatic prototypes, robots with laser-following features offer enhanced precision, adaptability, simplified guidance, object tracking, and research opportunities due to their suitability for multiple applications. Thereby, this paper explores the design and implementation of an Autonomous Aquatic Robot, capable of following a laser beam through an arrange of multiple RGB sensors feeding an embedded Artificial Neural Network (ANN), optimally trained through a metaheuristic algorithm (Earthquake Optimization Algorithm) to create a laser-following robot. Experimental results validate how Artificial Intelligence (AI) can be applied to generate a control structure for a laser-following robot, with over 99% of accuracy to generate activation signals by the laser presence detection, to provide a reliable signal for the autonomous prototype.