Controle de um manipulador robótico planar em série com redundância no espaço das tarefas

Abstract

The use of industrial robotic manipulators has gained significant prominence in modern times, and is also a major bet for the future of industries. Its use includes casting, welding and painting, load handling, product inspection, and displacement, among many others. Despite being widely used already, an extensive study continues to be carried out so that their tasks are executed in an optimized manner. To optimize performance, degrees of freedom can be used, which determine the mobility of the robot and, when in excess, result in kinematic redundancy. Redundancy resolution is the process of defining the joint positions among the infinite possible solutions. There are two types of resolution: local and global. This work used local resolution, which calculates the cost function gradient for all time instances to find the best joint configuration. The results were compared with a non-redundant manipulator, and both had a PID control. The manipulator was modeled using the Simscape Multibody package, and supporting functions were built in MATLAB. The results obtained show that the redundant manipulator had lower final effector position errors as well as significant reduction in joint torques.