A Method for Diagnosing a Robotic Complex Using Logical Neural Networks Apparatus

Abstract

An intelligent method for diagnosing the state of a robotic complex consisting of many actively interacting subsystems should not only have a set of logical rules established in practice, but also have the ability to generalize and classify input information, i.e. identify implicit (hidden) patterns. A logical neural network built on the basis of a variable-valued logical function meets the requirements for the diagnostic system. In this paper, the properties of variable-valued logical functions are investigated and an algorithm for constructing a logical neural network corresponding to the logical structure of these functions is proposed. It is shown that the logical neural network constructed by the proposed method not only preserves the original system of rules, but also has a number of additional useful properties for the logical processing of input information.