This paper presents the usage of holonomic mobile humanoid manipulators to carry out autonomous tasks in industrial environments, according to the smart factory concept and the Industry 4.0 philosophy. The problem of transporting lengthy objects, taking into account mechanical limitations, the conditions for avoiding collisions, as well as the dexterity of the manipulator arms was considered. The primary problem was divided into three phases, leading to three different types of robotic tasks. In the proposed approach, the pseudoinverse Jacobian method at the acceleration level to solve each of the tasks was used. The redundant degrees of freedom were used to satisfy secondary objectives such as robot kinetic energy, the maximization of the manipulability measure, and the fulfillment mechanical and collision-avoidance limitations. A computer example involving a mobile humanoid manipulator, operating in an industrial environment, illustrated the effectiveness of the proposed method.
Geometric Reasoning enabled One Shot Learning for Robotic Tasks
