The robot null spacenew uses for new robotic systems

Resumen

This doctoral thesis deals with the use of the robot redundancy to execute several tasks simultaneously at different levels of priority and its application to two different robotic systems. In particular, new uses of the robotic null-space are proposed. One of the robotic systems is studied in this thesis. The thesis memory is divided in two parts, according to the involved robotic system. The first robotic system is a humanoid robot on which a novel use of the null-space is presented: its use to transmit emotions to the users as a low priority task while the robot is executing a task with higher priority. A mapping to translate emotions as points in a three dimensional space to motions of the robot to be fed to the null-space is proposed. The proposed model is implemented and verified in a Pepper robot through a user study. In this study, a salutation motion is executed as the highest priority task, and the null-space is used to generate emotional motions. The analysis of the user study shows that the null-space is a proper mean to convey emotions to the user. The second robotic system is a novelty in itself: it consists of a teleoperation robotic system composed of a mobile manipulator, a UAV with a camera, and a haptic device. The operator teleoperates the mobile manipulator through the haptic device while getting visual feedback from the camera of the drone. The kinematic model of the robotic system, and an algorithm to coordinate the robot and the UAV so that the operator is released from the explicit command of the drone, are proposed. The results of a user study show that the proposed algorithm is appropriate in terms of the operator workload and the learning curve. A null-space algorithm is proposed to avoid own-occlusions of the robot end-effector, that is, situations where parts of the robot occlude the view from the camera of the object at the end-effector. An algorithm for the activation of the camera movements that guarantees that the operator does not lose sight of the teleoperated object, and which is only activated when the robot is about to leave the camera field of view, is presented. Additionally, the continuous inverse, a mathematical operator well suited for smooth transitions between the different levels of priority, is studied; two enhancements are proposed to avoid instabilities and that the lower priority tasks affect the higher priority tasks during transitions. The proposed algorithms are verified in simulations and in real experiments.