ML & Robotics Integrated with AUVs for Sub-Aquatic Applications.

Abstract

In the modern era of ground robots and flying Drones, the importance of Autonomous Underwater Vehicles (AUVs) cannot go remiss. Through customized sensor package either tethered to a Remote Operated Vehicle or onboard, an AUV is able to conduct tests, surveil, identify and monitor the underwater environment. Machine Learning (ML) and Deep learning with a vast array of algorithms have led to cutting edge results both dynamic and in real-time to accurately represent the underwater environment. To begin with, this paper will present the theoretical choices available for the control architecture which dictates the extent of autonomy of an AUV, followed by application-specific requirements. We will also touch upon the utility of AUVs and robotics for possibilities of underwater human habitation as an alternative to address the increasing scarcity of land. It has also had a tremendous impact in the naval sector, particularly coastal patrolling and safeguarding territorial waters from nefarious activities. The various issues concerning AUVs and robotics, drawn from the research work of researchers and scientists, that are an irritant across applications are also discussed. Notwithstanding, ML with its application-specific and groundbreaking algorithms has given enormous processing power. Therefore, this paper purports to examine how AUVs have made underwater terrain mapping, mineral exploration, disaster management, conservation and restoration of endangered coral reefs possible with increased efficacy and reduced time, thereby becoming the cornerstone reference for further research in this field.