Exploiting Redundancy in Underwater Vehicle-Manipulator Systems

The current work focuses on the development of a comprehensive scheme for the coordinated control of remotely operated vehicle–manipulator (ROVM) systems, and it proposes a novel mode for their operation. The proposed scheme consists of 2 main stages: redundancy resolution, and and robust model based control. In the redundancy resolution stage, the end-effector input commanded by a human pilot is distributed over the vehicle and manipulator. The redundant degrees of freedom (DOF) are used to accomplish secondary objectives. To this end, the Gradient Projection Method (GPM) is merged with a fuzzy logic based weighting scheme. Regarding the control of the system, a dynamic model is derived using the energy-based quasi–Lagrange approach. As opposed to a classic Lagrangian derivation, the quasi–Lagrange approach generates the equations in terms of body-fixed frames. As well, an adaptive sliding mode controller is implemented that constantly compensates for unknown dynamics throughout the vehicle-manipulator system. To demonstrate the efficacy of the scheme, a numerical case study is performed. Results illustrate that a complex end-effector spatial maneuver defined by a single 6-DOF pilot input can be accomplished with a 4-DOF manipulator mounted on a small ROV.