A state communication and software switching module and thin middleware layer for reconfiguration management in reconfigurable manufacturing systems.

Abstract

Reconfigurable Manufacturing Systems are a new area of operations and manufacturing research. The global need for production systems which can react rapidly to dynamic markets has increased in the last decade and will continue to drive changes in the manufacturing industry. The further development of RMS technologies is therefore highly important for future industries. The Reconfiguration Management and Middleware System (RMMS) developed in this research aimed to form a hardware-supported middleware technology which allows for the fast and seamless ramp-up of heterogeneous machine controllers on a newly reconfigured factory floor. The goal was to allow for the autonomous assignment and switching of software routines on machine controllers after a physical reconfiguration, thereby speeding up the ramp-up of the system. The technology was based on a recorded literature review and fits into the paradigm of RMS. The RMMS was developed not as a traditional software-heavy layer, but as a thin layer of software assisted by interactive mechatronic hardware, designed to remove heterogeneity in the control software. The system design was based on research into areas of engineering and operations management and followed the Mechatronic design approach. The literature led to a technology that takes the entire RMS paradigm into account and the development was conducted in conjunction with experiments to verify the individual functionality of each sub-system and ensure the overall system’s success. The RMMS uses hardware to handle heterogeneity and uses a positioning system (developed by the author) along with an intelligent processing system (a clustering algorithm and artificial intelligence engine) to construct data into a factory floor model. The positioning system, when assisted by the intelligence, operates at an accuracy of over 90%, which is comparable to commercial positioning techniques which cost over ten times more. The RMMS used the developed model to, autonomously and wirelessly, assign new programs to machine controllers after a physical reconfiguration, to complete a factory reconfiguration. The system was verified through practical scenarios constructed in the Mechatronics laboratory. Realistic reconfiguration operations were performed and the RMMS was required to detect changes in the factory floor and respond by assigning new, appropriate, software routines to each machine controller in the system. Experiments have proved that the system was capable of re-establishing operations in under half an hour, as opposed to a full day using manual techniques. The system has accurately switched between control routines based on the physical state of the factory floor, which amounts to control reconfiguration. The reconfiguration of factory floor control was successful in four out of four factory layouts tested and therefore successfully does a job no commercially available system can do.