MACHINE LEARNING DEVICE LEARNING OPTIMAL OBJECT HOLDING PATH, AND MACHINE LEARNING METHOD

摘要

PROBLEM TO BE SOLVED: To provide a robot with a hand having a function to hold multiple objects capable of minimizing a cycle time for storing the multiple objects into a container.SOLUTION: A machine learning device of the invention is a machine learning device that learns operational conditions for a robot to store multiple objects arranged on a transfer device using a hand for holding the multiple objects, and comprises: a state monitor part for monitoring state variables including at least one of positions/attitudes of the multiple objects, cycle time taken for storing the multiple objects into the container, and torque and vibration generated when the robot holds the object; a determination data acquisition part for acquiring determination data for determining margins with respect to respective allowable values for the cycle time, torque and vibration; and a learning part for learning the operational conditions of the robot according to a training data set constituted of a combination of the state variables and the determination data.SELECTED DRAWING: Figure 2