Dynamic positioning of remotely operated underwater vehicles

Remotely operated underwater vehicles (ROVs) play an important role in underwater technology. ROVs have been used for the past three decades in the off-shore oil industry. They are essential at depths were human diving is impractical. This is the case of the Brazilian Marlim oil field, which has water wells deeper than 1500 m. ROVs are used for constructing production production facilities, for inspection, and for investigation. ROVs are linked to a surface vessel by a tether cable that powers the ROV and transmits command signals and video images. Human pilots, aboard the surface vessel, usually operate the ROV through joysticks that command ROV thrusters. Video cameras and other sensors (e.g., depth-meter and compass) are the usual on-board navigation instruments. Acoustic systems (long baseline and short baseline) are often employed as auxiliary navigation systems. In many operations it is required that the ROV keeps its position close to some underwater structure (station keeping). This is usual in intervention operations involving ROV manipulators (e.g., opening a valve). Those operations are time consuming, difficult and tedious for the ROV pilot. The multivariable (six degree-of-freedom) and nonlinear nature of the ROV dynamics makes manual station keeping difficult. The pilot must continuously correct the ROV position to counteract the water current and the reaction forces caused by the motion of the manipulator. In this article we describe an automatic dynamic positioning system for ROVs (DPSROV) that is based on a mechanical passive arm (PA) measurement system. The PA system was selected from a group of candidate measurement systems, including long baseline, short baseline, inertial, speed logs, and taut-wire. The selection was based on several relevant criteria; namely, precision, construction cost, and operational facilities. We describe the DPSROV hardware and software; outline the dynamic model of the ROV and the position control system; discuss a conventional P-PI linear controller and the variable structure model-reference adaptive controller (VS-MRAC), which were implemented in order to show that the DPSROV can execute control algorithms of such diverse complexities and resource requirement; and present some experimental results obtained with the DPSROV. In a comparative study [1], we have concluded that the PA is the most reliable system for precise ROV dynamic positioning in a short-range workspace (few meters).