Motion simulator of an underwater drone for the study of maneuverability

Man's first attempts to build a submarine can be placed in 440 BC, however the first true marine vehicle with engine propulsion and capable of diving was created in Spain in 1867 and could dive up to 30 meters. Currently modern models reach much higher depths (about 400 meters). However, in order to overcome these limitations, for example the space to be dedicated to the crew and the oxygen production plants, they are opting to produce drone submarines. Drones would allow the achievement of depth and greater autonomy, moreover, maintaining smaller dimensions that would allow better maneuverability This thesis aims to develop a control system able to drive a submarine drone in several maneuvers meeting specific requirements. The hypothesized vehicle uses a single big propeller at the stern maneuvered from a rudder system composed by a X configuration at the stern and a two horizontal rudderes in the bow. The first chapter deals with the mathematical formulation describing the 6-degree dynamics of a drone in water, while in the final phase the topic becomes the develop of an automatic control system and its tuning based on a genetic algorithm. Several maneuvers have been analyzed to validate the capability of the genetic algorithm to identify the control system parameters that maximize the maneuver's performance indicators. The work was supervised with Cetena, a naval engineering company that is developing this software for the control and design of experimental drone submarines.