The Training Set Automating for a Neural Network Model of an Industrial Robot Control System

The Training Set Automating for a Neural Network Model of an Industrial Robot Control System

연구 분야: Infrastructure

학회: 2024 International Conference "Quality Management, Transport and Information Security, Information Technologies" (QM&TIS&IT)

The industrial robot control system solves a large number of tasks that affect the ease of integration and efficiency of the manipulator in the technological process. One of its main tasks is to calculate in real time the trajectory of the manipulator flange movement taking into account the selected type of movement. In particular, this task can be solved using a neural network model. The article describes the solution of the problem of automating the creation of a training set for a neural network model that forms the trajectory of movement of an industrial manipulator. The source of data for the set is a KUKA KR4 R600 manipulator and its digital twin created in the “Rational Production” software. The training set includes many sets of the manipulator joint angle vectors (generalised coordinates) that uniquely define the current position of the robot and its flange in space.