Expert System for Discrete Component Digital Circuit Design

摘要

The design of digital circuits is cumbersome and prone to errors. Current estimates show that a circuit designed with 5 integrated circuits has less than a 50% chance of working, due to wiring errors, when power is first applied. As the complexity increases, the probability of the circuit working on the first try decreases rapidly. Design errors fall into two categories: wiring or logic errors. Wiring errors consist of improperly connected gates, missing connections, and violations of fanout, or race conditions. Logic errors, which are more complex to detect, require knowledge of intended circuit function. This thesis designs, develops, and tests an expert system which decomposes digital circuits into subproblems described by database technology in order to detect wiring errors. Information needed to connect chips together is viewed as knowledge base information for the expert system. Information such as number of pins, value of each pin (input, output), type of chip, etc., are represented in a database.