A knowledge-based control paradigm for real-time systems

Abstract

The authors examine the application of knowledge-based symbolic control to the management of execution and configuration of a complex numerical control system. Symbolic processing is used to implement inference of system state and internal communication for inference and control. The flavor system provides an object-oriented programming environment in which the inference engine and knowledge base for the symbolic controller are realized. System communication is accomplished by asynchronous message passing using a mailbox facility. The particular target application considered is a software-intensive radio, which is envisioned as being digitally implemented. Symbolic processing is used to internally control the radio down to the module level. Testing is via computer emulation (Monte Carlo).<>