A Refinement-Based Approach to Deriving Train Controllers

The purpose of this paper is to demonstrate how transformation can be used to derive a high integrity implementation of a train controller from an algorithmic specification. The paper begins with a general discussion of high consequence systems (e.g., software systems) and describes how rewrite-based transformation systems can be used in the development of such systems. We then discuss how such transformations can be used to derive a high assurance controller for the Bay Area Rapid Transit (BART) system from an algorithmic specification.