Symbolic diagnosis for intelligent control

Abstract

The results of research intended to create a symbolic diagnostician to support intelligent control of numerical processors and/or processes are reported. Example applications include real-time signal processors, industrial automation, and aerospace power systems. The approach is to create a generic, symbolic inference engine to interpret data from real-time numerical processes. The interpreted data are then utilized by companion symbolic and numeric modules resulting in a dynamic, intelligent real-time control architecture. General results are obtained while focusing research efforts on an initial target application-a software-intensive radio receiver/processor. Object-oriented programming techniques are used due to ease of knowledge engineering and potential parallels to hardware implementation.<>