Publié : 27 février 2019
The Chair of Control Engineering at the University of Augsburg is developing the Modular Cable Robot (MoCaRo). In the future, it will operate as a flexible handling system that can be quickly installed in a production facility, for example, but it is still highly dynamic and precise. This requires new calibration and control algorithms based on a dynamic model of parallel kinematics.
The demonstrator with the robot suspended from the sides.
The demonstrator consists of an aluminum profile with cables tensioned upwards that can move an end effector in the room. The cable is guided by a cable drum and a deflection pulley.
In the demonstrator, the robot moves freely within a cube of 2.5 m edge length.
The working area of the robot consists of a cube with an edge length of 2.5 m. The ropes are driven by Beckhoff servo motors.
A SCALEXIO LabBox (pictured below) is used as the SCALEXIO system. It includes the SCALEXIO Fieldbus Solution with an EtherCAT interface and a DS6101 Multi-I/O Board for signal conditioning. The SCALEXIO Processing Unit (top) serves as an external, powerful computing unit.
The motors for moving the robot are controlled via the EtherCAT interface of a dSPACE SCALEXIO system. By using an EtherCAT master within the SCALEXIO system, it is possible to control both the servo motors and the digital and analog inputs and outputs in real time. To implement real-time control of the cables in the micrometer range, it is necessary to evaluate 32-bit encoder signals for each motor so that the motor can be controlled on the basis of torque.
The controller model was developed in MATLAB®/Simulink®. The control results were evaluated and visualized in dSPACE ControlDesk. In addition, the measured values can be recorded automatically via an XIL API interface, for example, to set the cable length in MATLAB® by means of scripts or to measure the end effector position. The smooth interaction of hardware and software ensures fast rapid control prototyping. Adjustments in the Simulink® model can be loaded to the system and executed at the click of a button. Evaluation with ControlDesk allows drawing conclusions that can lead to adjustments in the Simulink® model. In addition, other bus-based hardware components can be flexibly added to the system to respond to new development tasks.
By embedding the lightweight rope robot in an active sensor and control system based on the SCALEXIO real-time system, MoCaRo provides a precise, dynamic and easily expandable overall system.
M. Sc. Marcus Hamann is researching the development of robot controllers at the Chair of Control Engineering of the Institute of Software & Systems Engineering at the University of Augsburg, Germany.