Study Of Translation Manipulation Of A Dexterous Robot Hand In Soft Contact Environment With Moving Contact Point Model

Dexterous manipulation(in-hand manipulation) is the key function of a dexterous robot hand, and much attention has been paid to this area. And a translation manipulation under soft contact model is chosen as the center of this research project, achieve the translation manipulation and slippage control during the manipulation is expected.Get the information of object’s position and the contact force is the foundation of achieving the translation manipulation. The DLR/HIT II dexterous robot hand is used in this project, and a distance sensor and force sensor is added to the electrical system, as well as the relative circuit to process the signals and the model to analyze and transmit the signals.The phenomenon of contact point moving during a dexterous manipulation is studied during this research. We get the relationship between the contact point and the rotation of the fingertip through the analysis of the fingertip’s profile curve and the object’s profile curve without the use of a tactile sensor.Build a proper model of the manipulation process will provide meaningful guide to the manipulation. In the manipulation of soft contact, Hertz contact model is chosen to get the relationship between contact force and the depth of contact as well as the contact area. A coulomb amendment model is chosen to describe the friction model of soft contact. And the dynamics of the robot finger and the manipulated object is analyzed as well.Trajectory planning is the key to achieving the dexterous manipulation. Based on the models has been built above the trajectory planning is studied in this research, trajectory generating algorithms are developed to achieve fixed contact force manipulation of different contact point models, optimized contact force manipulation, and a trajectory optimizing algorithm is provided as well with the goal of achieving slippage control during the manipulation.As last, an experiment platform is built, and the effectiveness of the model of contact point, the trajectory generating algorithm and the trajectory optimizing algorithm are confirmed through different experiments in this platform.