Posture optimization in robot-assisted machining operations

This paper aims to provide a robot performance index to evaluate the effectiveness of the actuator torque and joint rates in producing a prescribed robot gesture during a machining operation. Most performance indices proposed in the literature are only posture-dependent; however, to properly evaluate the goodness of the force transmission when the robot performs a task, the index should be also task-dependent. The index proposed here, of the kinetostatic type, targets applications that call for optimizing the posture of a redundant robot. In this light, dynamic effects are not relevant to this study. The concept is formulated rigorously for the architecture most commonly found in industrial robots of the serial type, i.e., composed of six revolute joints. Furthermore, a simplified version of the same index is obtained for robots of the decoupled type, while neglecting the end-effector angular velocity and the torque applied by the cutting force. Finally, the simplified index is used to optimize the posture of a six-revolute robot while performing a five-axis machining task, thus ending up with a redundant robot for the given task.

► The postures of a redundant six-revolute industrial serial robot were optimized for the robot transmission ratio (RTR).
► A general definition of the RTR for n-dof serial robots is presented which is dimensionless.
► The RTR is defined in the joint space and is not only posture-dependent but also task-dependent.
► The RTR is formulated rigorously for a general serial robot by means of screw theory and simplified for decoupled robots.
► The RTR is maximized for a redundant robot over a five-axis task using a gradient map.