A formal methodology for constructing decoupled manufacturing system control software components.

Abstract

There is an increasing need for flexibility in manufacturing systems. Achieving flexibility in a cost effective way, directly impacts the hardware and software used to control the system. One way to reduce this impact is through reuse. To increase the opportunity for reuse, we propose a control software development methodology which allows control software to be partitioned into components, based on their likeliness to be modified. Our approach consists of two parts, a software architecture for reuse and a modeling formalism. The architecture decouples the control software into components based on their likeliness to be modified. We employ software/hardware components, a modern software engineering tool, to design and construct portions of this decoupled component architecture. The modeling formalism is based on finite automata. Finite automata have an enormous base of established theoretical and practical results, and are easily used to represent the dynamics and control of manufacturing systems, as well as many other Discrete Event Dynamical Systems (DEDS). We use the formalism in specifying, designing, and constructing portions of the control software. In particular, the formalism helps us to understand the dynamical behavior of the system, the way this behavior can be modified, and the way the system's entities interact with each other. It also helps in expressing the tasks to be implemented by the control software. Our contribution builds on earlier results by including comprehensive representations for concurrency and control information, and by emphasizing the duality between automata and formal language representations. In addition, we provide specific constructs for capturing interactions between components of the system. We demonstrate the utility of our approach in the construction of control software for a typical manufacturing system application. In this application we are able to illustrate all of the aspects of our modeling formalism, and show how the software architecture and modeling formalism are used together to improve reuse.