Software modeling for reconfigurable machine tool controllers.

Reconfiguration of hard real-time software for machine tool controllers requires an architectural framework that supports a range of potential configurations which are not specifically identifiable in advance. However, at present no such framework exists, nor is there a well-established process to design it. This research investigates an incremental, iterative process to develop a requirements specification model and a domain-specific architecture to facilitate the integration of software for hard real-time machine tool control functions. It is an early stage study in a very complex and difficult subject of long-range research. The research is performed as a series of experimental steps in which the domain model is developed incrementally and a test application is prototyped at each step to evaluate the model and discover limitations. Performance feasibility is partially validated through multi-axis motion control tests on a machine tool. Results include a set of domain-specific heuristic principles, procedures, rules, and constraints on task interaction patterns, and a set of reusable object classes to develop a configuration of hard real-time tasks that integrate continuous process control with discrete events, e.g., intervention of a tool breakage signal to stop servo-controlled motion. Contributions in the domain modeling process include a process of generalization from a specific case, a technique to bound domain, and a technique to make it extensible. The architecture is extensible through specialization of its component classes, through the aggregation of components into subassemblies, and through the development of progressively relaxed constraints on interaction patterns for subdomains requiring less stringent timing constraints. The model is reusable across various types of automated functions for material processing, inspection, and material handling, thus facilitating their computer integration.Ph.D.Applied SciencesComputer scienceMechanical engineeringSystems scienceUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/130202/2/9721947.pd