DEPTH POSITIONING CONTROL OF A SPHERICAL UNDERWATER ROBOT VEHICLE (URV)

Underwater Robot Vehicles (URVs) are an underwater system that has its own power and controlled by an onboard computer. URV has been deployed to the numerous duties without human interventions, such as repairing subsea systems and installing networks for scientific research and data collection. In most of these applications, a stable motion is required for the URV. However, the existing URV designs with the fixed ballast tank which generates an internal dynamic for the water inside the tank, leading to the unstable motion for the URV. It hypothesized that an alternative tank mechanism which has variable ballast can perform better and improve the stability of the URV motion. This work focuses on developing a spherical URV has a variable ballast tank mechanism. The complete modeling system of the URV was built and the dimensions and control system components were defined using Simulink/MATLAB and the characteristic of the system was analyzed using the open loop response. A numerical simulation was conducted by using three different controllers namely, PID, Sugeno - FIS and Mandani - FIS. These controllers were numerically evaluated in both fresh and seawater conditions with ranges of depth position up to one thousand meters. In terms of a fast response, Mandani - FIS showed a fast response to control the depth position of the URV without overshooting. To validate the FLC based on Mamdani - FIS, an experimental work was conducted, in which the spherical URV was tested for both downward and upward trajectories in one meter freshwater. Velocities of 0.106 m/s and 0.095 m/s resulted respectively, for downward and upward trajectories. A comparison between the experimental and numerical simulation showed a difference of time response for both downward and upward trajectories. This difference is due to the effect of water pressure, which increases with increasing depth, as well as the effect of gravity. Additionally, the URV motion was affected by the horizontal waves associated in the testing pool when the experimental work was conducted, but for numerical simulation, the effect of wave load was not considered