Requirements specification for a real-time embedded expert system for rapid prototyping

Abstract

Several commercial expert system shells provide knowledge engineers with the capability of developing expert system applications, but are not able to meet the size constraints or provide the run-time performance needed to address problems associated with delivery of embedded real-time applications. In addition, there is no provision to provide deliverable code in Ada, a requirement for many US DoD systems. The embedded consultant project addresses these issues by knowledge-base size reduction by means of code optimization techniques, and by an inference engine, written in Ada, designed for real-time applications. A new rapid prototyping approach for real-time applications is described. The methodology is based on the reuse of existing software components. The approach, and the significance of the work are discussed.<>