An expert system for circuit board repair

Abstract

Using a hierarchical approach, an expert system for performing circuit board repair tasks within a robotic workcell has been developed. The new circuit board repair system is rule-based and includes a nonmonotonic inference engine using forward chaining and nonblocking functions, a database describing a robotic workcell, and rule sets into which repair tasks are implemented. Parallel-processing capability is achieved by using nonblocking functions. Action synthesis is based on matching rules to the given database. The system was programmed in Common LISP on a SUN 3 workstation under UNIX BSD version 4.3 and on a dedicated Symbolics 3600 LISP machine. The development, implementation, and testing of the circuit board repair system are described.<>