IIC Journal of Innovation 12th Edition | Page 42

Digital Twin Development for Serial Manipulators: Data Driven Optimized Planning and Sequencing of Tasks conference paper by Huynh et al. 19 The real- time cobot joint data from the cobot asset are streamed to the cobot DT model for live monitoring and synchronized cobot movements. the shaft assembly station. Next, the engineer examines the actual cobot with a new path as a final test and releases for shaft sub-assembly productions. That will trigger the warehouse system to start delivering the necessary parts to the shaft assembly station by an AGV. Subsequently, the shaft assembly station operator starts the customized work order at MES, and the cobot assembles the sub-assemblies accordingly. The completed products are moved to the warehouse by the AGV and the smart inventory system informs ERP for the order completion. The robotic engineer starts working on creating the new cobot motion path virtually once he is notified about the customized work order from the manager. This process will not interrupt any operations on the on- going production line. There can be many iterations and validation processes carried out for developed simulation models, and the virtual cobot is useful for reviewing the new cobot path planning. As for the preparation of the customized work order, the deep learning model of assembly parts recognition is trained with computer vision techniques at the same time. Later, the trained part recognition model will be tested and integrated with the new motion plan for picking, placing and assembling accordingly for the shaft assembly process. The result of motion planning software is transferred to the cobot DT model (virtual cobot) for verification of correct cobot paths by the engineer. This is where the data is transformed into information used for effective decision-making. After confirmation, the motion plan is released through the information gateway to the robot controller. In this use case, the communication and information flow between the associated assets on the shop floor are critical to realize the digital twin technology. With those capabilities, the monitoring and control of the production process and accessing enterprise information are possible for the production manager to make decisions in the MICR. Apprehending the resource requirements and planning necessary tasks are the initial steps taken by the manager to complete a customized order on time. Hence, once the cobot at the shaft assembly station is ready to perform the customized work order, this state triggers to allow that order to start in MES. That order is scheduled to proceed next at the assembly line in the MES interface as soon as the on-going work order is finished. The vertical integration of process, asset data and enterprise information at the IoT platform enables this operation. Once that work order is started, the appropriate raw materials are collected Only after the verification of the new cobot path is completed in the virtual world, the final motion path is laid out into the robot controller which connects with a cobot at 19 Huynh, B. H., Akhtar, H., & Myo, K. S. (2019). A Universal Methodology to Create Digital Twins for Serial and Parallel Manipulators, 2019 IEEE International Conference on Systems, Man, and Cybernetics, IEEE, Accepted. IIC Journal of Innovation - 37 -