Complete System for Water Monitoring Using Unmaned Surface Vehicles and Deep Learning
Aquatic Monitoring, Autonomous Robotics, USV, Deep Learning
This paper proposes an intelligent robotic water monitoring system, called SRIMA, which uses concepts and techniques of mobile robotics and machine learning, with the objective of reducing the costs of monitoring, proposing a new metric of water quality index and to anticipate emergency situations, in order to mitigate or even prevent the worsening of the current water situation. Basically, four unmanned surface robotic vehicles monitor the reservoir area based on path planning algorithms, from four points defined at the start of the mission by the base station. For each starting point, marginal points are generated in order to maximize the collection area. The vessels monitor the defined area for a set time, return the base station and send the collected data of pH, temperature, conductivity, salinity, dissolved oxygen, oxidation potential and total solids. In the base station, artificial intelligence and image processing algorithms generate the representation map of the monitored area. It is intended to obtain the complete mapping of the reservoir under analysis, so as to predict abnormality at the collection points, if any, and indicate which situation may be causing this problem.