• ottimizzazione
• modellistica
• controllo

This thesis will presents the state of the art and the personal contribution to the control-oriented modeling and control of marine vehicles. It is divided in two parts, the first one describes the scientific results related to the modeling and control aspects, while the second part describes an application related to the UAN project and a real-time distributed supervision system.
The first result, described in chapter n.1, was motivated by the existence of a new class of hybrid underwater vehicles equipped with classic AUV and glider actuators. In particular, a dynamic model was derived for the general class of hybrid-propulsion vehicles, to which the case study eFolaga belongs. This development was made possible by considering a kinematic relaxation: the effects of the Centre of Gravity (COG) variations on the linear speed of the vehicle were neglected. The resulting simplified kinetic model, including the COG motion variation effects, takes a standard control-oriented form and consequently it can be easily used with all the available literature results regarding control and related aspects. This simplification works out if the COG speed is very small as compared to the vehicle speed, as it is the case of the eFolaga and of most vehicles with internal mass displacement systems.
The second result, described in chapter n.2, was motivated by the need of a underwater vehicle capable to achieve high-energy efficiency by extracting propulsion directly from the sea-waves. The novel vehicle was designed to combine typical underwater capabilities such as underwater navigation and low-consumption gliding, with the clean-energy oriented feature of sea-wave energy extraction and propulsion conversion. The resulting vehicle, named Underwater Wave Glider (UWG) exploits the sea wave potential and the surface fluid profile with hydrodynamic wings to achieve a clean-energy passive propulsion. Currently the UWG is object of a combined study between ISME1 and CSSN2 .
The importance of analyzing systems under constant sea-wave excitation motivated the study described in chapter n.3. In particular a method to combine surface models, described for instance with Motion Response Amplitude Operation (RAO), with classic underwater model was developed to evaluate system performances during floating motion. The resulting hybrid model, while capturing frequency domain specification of surface motion model, is able to provide time-varying access to the sea-wave induced effects acting on submerged vehicle parts.
The chapter n.4, concluding the first part of the thesis, is dedicated to the control-allocation problem. In fact, the capability to reproduce coherently forces and moments through an intelligent use of actuators is a common requirement for marine vehicles control systems. The results presented, applied to a quite diffused class of marine vessels, can also be extended to other underwater or surface vehicle classes, by including the proper actuator spatial distribution and limitations. In particular, a parametric sequential quadratic programming method is proposed to solve the problem of control allocation with unconstrained forces/moments references. The new formulation, as shown in detail in the chapter, improves performances, with respect to the state-of-the-art, in case of not-feasibility or actuator saturation conditions. The second method, is used to properly distribute actuators configurations by exploiting a norm-infinity bounded reference set. This particular case has a direct impact on all vehicles of the considered class, equipped with human-interface-device (HID) systems. The implication of these new methods, together with the new
drive-by-wire methodology, could affect considerably the nowadays vehicles maneuvering capabilities.
The second part of the thesis is dedicated to the applications. In particular chapter n.5 will discuss experimental results of the UAN project final experiment in Throndheim, Norway (2011). Finally, chapter n.6 will present a cross-platform distributed system, named DCL, used to implement Hardware-In-the-Loop (HIL) and Software-In-the-Loop methodologies and to supervise real-time RTAI processes.