Planning and Motion Control for Underwater Bimanual Soft Manipulator in Underwater Grasping Task

Researchers have always been curious about underwater biodiversity and energy diversity. However, autonomous intelligent and friendly manipulation in the restricted aquatic environment remains challenging. Inspired by the friendly bimanual manipulation of humans in the restricted aquatic environment, this work proposes an underwater bimanual soft manipulator (UBSM) system to complete the grasping task. Since an underwater object may move with flow during grasping, this work designs a hierarchical coordination framework consisting of an upper task planning level and a lower motion control level to integrate UBSM perception, interpretation of object state, task state, and UBSM actions. An upper planner at the task planning level is designed to realize a perception-action strategy consisting of two bimanual coordination modes. And an online-learning-inverse-kinematics controller at the motion control level is proposed to enable the UBSM to handle the randomness and adapt to the underwater environment. We validate the effectiveness of the designed upper planner and proposed motion controller in the underwater grasping task by experiments.