A Receding Horizon Motion Planner for Underwater Vehicle Manipulator Systems

摘要

In this paper, a receding horizon (RH) motion planner is developed for an underwater vehicle manipulator system (UVMS) consisting of an unmanned underwater vehicle (UUV) and a robotic arm with multiple joints. The desired task is assigned as driving the end effector to a predefined pose while keeping roll and pitch of the UUV in a small range. Model-based constraints such as input saturations and mechanical limits are explicitly considered within the RH framework, and thus the generated velocities are executable for the real system. Compared with the task priority control approach, which has been widely studied for motion planning of UVMSs, the RH planner is better in its optimality and capability to deal with complicated environments and mission requirements. Performance of these two approaches is compared in simulation studies and effectiveness of the proposed RH motion planner is verified.