Method and apparatus for a fuzzy self-adaptive control system

摘要

To resolve a problem of the repairing of an electronic system, a machine will undergo a learning stage during the replacement of a part of the machine. This learning stage can be monitored manually or automatically on a limited range of variation. During this learning stage, readings are taken of the resultant modifications on a structure of the membership functions and of interference rules, memorized in a fuzzy logic program that controls the machine. These membership functions and/or theses rules are then modified as a function of this results. It is shown that it is possible, in this way, to repair a machine without replacing a defective part by a strictly identical part and without having to bring the machine back to the laboratory.