Study of innovative autonomous marine vehicles for monitoring in remote areas and shallow waters The Shallow Water Autonomous Multipurpose Platform (SWAMP)

The main objective of the research activity covered by the present thesis is the design of an Autonomous Surface Vehicle for the monitoring of environmental areas characterised by shallow water, difficult access and harsh environment, namely the Wetlands. Wetlands are those geographic areas where water meets the earth that cover between 5 and 8% of the Earth’s surface. Wetlands include mangrove zones, swamps, bogs and marshes, rivers and lakes, alluvial plains and flooded forests, shallow coasts and coral reefs. In recent years, their importance is becoming more and more recognized and various international conventions, directives and projects work on their protection. Their importance is related to the fact that these areas are essential ecosystems considered among the world’s most productive environment. Classified as natural purification systems and carbon resources for fauna, Wetlands provide the water and productivity upon which biological diversity relies for the growth of an enormous amount of species of plants and animals. Their importance is also related to human activities since Wetlands can be exploited commercially for fishing but especially become important when thinking that the protection of these areas can also help fighting the disasters resulting from human impact on the environment and its role in the climate change. The lack of a hydro-graphic vessels capable of performing shallow water measurements at depths of less than 1m has led to unreliable maps and data, thus motivating research on innovative technical approaches for executing the tasks of water sampling, limnological surveys, bathymetric analyses and monitoring of water quality. In recent years a variety of robotic approaches to improve the quality, speed, and accessibility of surveys have been explored by research groups using both commercial and ad-hoc solutions. In this thesis a prototype of Autonomous Surface Vehicle (ASV) named SWAMP (Shallow Water Autonomous Multipurpose Platform) is proposed as the base for an innovative class of reliable modular re-configurable lightweight ASVs for extremely shallow water applications. The vehicle was studied to solve the problem of monitoring the water status in the Wetlands but the SWAMP class ASV will also be able to support, as test-bed, the research on many aspects of marine engineering and robotics like propulsion issues, structure issues, artificial intelligence, cooperative distributed control, Guidance, Navigation and Control (GNC) systems as well as innovative technological solutions in terms of communication, materials, sensors and actuators. The heterogeneity of the themes treated by this thesis relies on the fact that the whole aspects of design were taken into consideration. In this thesis the description of the design, modeling, construction and testing of the new concept of Autonomous Surface Vehicles (ASV) is illustrated. The motivations behind the necessity of a new system are described in Chapter 1 and 2 while in Chapter 3 the general considerations on the requirements that led to the definition of the specifications of a special layout are reported. In Chapter 4 the design of the vehicle layout is illustrated together with the description of an innovative soft-material hull structure on which extensive analyses in towing tank were performed. The tests were carried out both in deep and shallow water to completely identify the surge motion of the ASV. The hardware, software and mechanical modularity represent some main ideas behind the conception of SWAMP. The two hulls of SWAMP are two separate modules. Each hull can be composed of more or less structural elements, actuation modules, powering elements, control units, sensors. This can be done without constraints thanks to the novel communication architecture all based on Wi-Fi modules. In Chapter 5 the thrusters expressly studied for environmental monitoring in the extremely shallow waters of the Wetlands (rivers, lakes swamps, marshes, mangroves..) are illustrated. These systems were modeled, designed and constructed on the Pump-Jet concept. Four Pump-Jet Module s for a class of small/medium size ASV were built and the description of the design and tests are reported in the thesis. The extremely modular hardware control system of SWAMP is described in Chapter 6 where also the modules composing the vehicle are described. Once assembled, the vehicle was tested at sea in various environments. A series of pioneering tests with the application of Machine Learning and the citizen’s engagement in teaching to a robot to self-control are described in the Chapter 7, together with more standard results. The algorithm for the training of a neural network for the control of SWAMP was tested also by using the simulator described in Chapter 8. The conclusions of this work, reported in Chapter 9, are correlated with a visionary analysis of the possible applications of SWAMP in a series of futuristic research trends of marine robotics.